* [lustre-devel] [PATCH 0/3] First set of Intel branch merger for libcfs/lnet
[not found] <uja.ornl@gmail.com>
@ 2015-05-21 22:46 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: lustre-devel
This is the first set of patches that are current in the Intel branch
which are apart of the effort to cleanup LNet/libcfs.
James Simmons (3):
staging:lustre: remove tcpip abstraction from libcfs
staging:lustre: remove kernel defines in userland headers
staging:lustre: cleanup libcfs lock handling
.../staging/lustre/include/linux/libcfs/libcfs.h | 21 -
.../lustre/include/linux/libcfs/libcfs_private.h | 51 --
.../lustre/include/linux/lnet/api-support.h | 44 --
drivers/staging/lustre/include/linux/lnet/api.h | 51 +-
.../staging/lustre/include/linux/lnet/lib-lnet.h | 264 ++-------
.../staging/lustre/include/linux/lnet/lib-types.h | 649 +++++++++-----------
.../lustre/include/linux/lnet/linux/api-support.h | 42 --
.../lustre/include/linux/lnet/linux/lib-lnet.h | 71 ---
.../lustre/include/linux/lnet/linux/lib-types.h | 45 --
.../staging/lustre/include/linux/lnet/linux/lnet.h | 56 --
.../lustre/include/linux/lnet/lnet-sysctl.h | 49 --
drivers/staging/lustre/include/linux/lnet/lnet.h | 11 +-
.../staging/lustre/include/linux/lnet/lnetctl.h | 36 +-
drivers/staging/lustre/include/linux/lnet/lnetst.h | 665 ++++++++++++--------
drivers/staging/lustre/include/linux/lnet/nidstr.h | 73 +++
drivers/staging/lustre/include/linux/lnet/ptllnd.h | 93 ---
.../lustre/include/linux/lnet/ptllnd_wire.h | 119 ----
.../staging/lustre/include/linux/lnet/socklnd.h | 91 ++--
drivers/staging/lustre/include/linux/lnet/types.h | 266 +++++++--
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 4 +-
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h | 9 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 16 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 3 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 7 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 21 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 30 +-
drivers/staging/lustre/lnet/lnet/Makefile | 7 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 47 +-
drivers/staging/lustre/lnet/lnet/api-ni.c | 100 +---
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-eq.c | 2 +-
drivers/staging/lustre/lnet/lnet/lib-md.c | 6 +-
drivers/staging/lustre/lnet/lnet/lib-me.c | 24 +-
drivers/staging/lustre/lnet/lnet/lib-msg.c | 22 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 616 ++++++++++++++++++
drivers/staging/lustre/lnet/lnet/module.c | 4 +-
drivers/staging/lustre/lnet/lnet/router.c | 145 -----
drivers/staging/lustre/lustre/include/lustre_net.h | 4 +-
drivers/staging/lustre/lustre/libcfs/Makefile | 1 -
drivers/staging/lustre/lustre/libcfs/fail.c | 2 +-
.../lustre/lustre/libcfs/linux/linux-tcpip.c | 623 ------------------
.../lustre/lustre/libcfs/linux/linux-tracefile.c | 4 +-
drivers/staging/lustre/lustre/libcfs/module.c | 11 +-
drivers/staging/lustre/lustre/libcfs/tracefile.c | 2 +-
44 files changed, 1856 insertions(+), 2572 deletions(-)
delete mode 100644 drivers/staging/lustre/include/linux/lnet/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
create mode 100644 drivers/staging/lustre/include/linux/lnet/nidstr.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
create mode 100644 drivers/staging/lustre/lnet/lnet/lib-socket.c
delete mode 100644 drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
^ permalink raw reply [flat|nested] 44+ messages in thread* [PATCH 0/3] First set of Intel branch merger for libcfs/lnet
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
This is the first set of patches that are current in the Intel branch
which are apart of the effort to cleanup LNet/libcfs.
James Simmons (3):
staging:lustre: remove tcpip abstraction from libcfs
staging:lustre: remove kernel defines in userland headers
staging:lustre: cleanup libcfs lock handling
.../staging/lustre/include/linux/libcfs/libcfs.h | 21 -
.../lustre/include/linux/libcfs/libcfs_private.h | 51 --
.../lustre/include/linux/lnet/api-support.h | 44 --
drivers/staging/lustre/include/linux/lnet/api.h | 51 +-
.../staging/lustre/include/linux/lnet/lib-lnet.h | 264 ++-------
.../staging/lustre/include/linux/lnet/lib-types.h | 649 +++++++++-----------
.../lustre/include/linux/lnet/linux/api-support.h | 42 --
.../lustre/include/linux/lnet/linux/lib-lnet.h | 71 ---
.../lustre/include/linux/lnet/linux/lib-types.h | 45 --
.../staging/lustre/include/linux/lnet/linux/lnet.h | 56 --
.../lustre/include/linux/lnet/lnet-sysctl.h | 49 --
drivers/staging/lustre/include/linux/lnet/lnet.h | 11 +-
.../staging/lustre/include/linux/lnet/lnetctl.h | 36 +-
drivers/staging/lustre/include/linux/lnet/lnetst.h | 665 ++++++++++++--------
drivers/staging/lustre/include/linux/lnet/nidstr.h | 73 +++
drivers/staging/lustre/include/linux/lnet/ptllnd.h | 93 ---
.../lustre/include/linux/lnet/ptllnd_wire.h | 119 ----
.../staging/lustre/include/linux/lnet/socklnd.h | 91 ++--
drivers/staging/lustre/include/linux/lnet/types.h | 266 +++++++--
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 4 +-
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h | 9 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 16 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 3 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 7 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 21 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 30 +-
drivers/staging/lustre/lnet/lnet/Makefile | 7 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 47 +-
drivers/staging/lustre/lnet/lnet/api-ni.c | 100 +---
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-eq.c | 2 +-
drivers/staging/lustre/lnet/lnet/lib-md.c | 6 +-
drivers/staging/lustre/lnet/lnet/lib-me.c | 24 +-
drivers/staging/lustre/lnet/lnet/lib-msg.c | 22 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 616 ++++++++++++++++++
drivers/staging/lustre/lnet/lnet/module.c | 4 +-
drivers/staging/lustre/lnet/lnet/router.c | 145 -----
drivers/staging/lustre/lustre/include/lustre_net.h | 4 +-
drivers/staging/lustre/lustre/libcfs/Makefile | 1 -
drivers/staging/lustre/lustre/libcfs/fail.c | 2 +-
.../lustre/lustre/libcfs/linux/linux-tcpip.c | 623 ------------------
.../lustre/lustre/libcfs/linux/linux-tracefile.c | 4 +-
drivers/staging/lustre/lustre/libcfs/module.c | 11 +-
drivers/staging/lustre/lustre/libcfs/tracefile.c | 2 +-
44 files changed, 1856 insertions(+), 2572 deletions(-)
delete mode 100644 drivers/staging/lustre/include/linux/lnet/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
create mode 100644 drivers/staging/lustre/include/linux/lnet/nidstr.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
create mode 100644 drivers/staging/lustre/lnet/lnet/lib-socket.c
delete mode 100644 drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: lustre-devel
From: James Simmons <uja.ornl@yahoo.com>
Since libcfs no longer builds for user land we can
move the TCPIP abstraction that exist to the LNET
layer which is the only place that uses it. Also
the migrated code will use native linux kernel
APIs directly instead of with wrappers.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13760
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 17 -
.../staging/lustre/include/linux/lnet/lib-lnet.h | 15 +
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 16 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 7 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 17 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 30 +-
drivers/staging/lustre/lnet/lnet/Makefile | 7 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 43 +-
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 616 +++++++++++++++++++
drivers/staging/lustre/lustre/libcfs/Makefile | 1 -
.../lustre/lustre/libcfs/linux/linux-tcpip.c | 623 --------------------
14 files changed, 701 insertions(+), 716 deletions(-)
create mode 100644 drivers/staging/lustre/lnet/lnet/lib-socket.c
delete mode 100644 drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 947df7e..f54bab6 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -88,23 +88,6 @@ static inline int __is_po2(unsigned long long val)
int libcfs_arch_init(void);
void libcfs_arch_cleanup(void);
-/* libcfs tcpip */
-int libcfs_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
-int libcfs_ipif_enumerate(char ***names);
-void libcfs_ipif_free_enumeration(char **names, int n);
-int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
-int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-void libcfs_sock_abort_accept(struct socket *sock);
-int libcfs_sock_connect(struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port);
-int libcfs_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
-int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
-int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
-int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
-int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-void libcfs_sock_release(struct socket *sock);
-
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
#define LNET_ACCEPTOR_MAX_RESERVED_PORT 1023
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index 0038d29..c00a2ca 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -846,6 +846,21 @@ int lnet_acceptor_port(void);
int lnet_acceptor_start(void);
void lnet_acceptor_stop(void);
+int lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
+int lnet_ipif_enumerate(char ***names);
+void lnet_ipif_free_enumeration(char **names, int n);
+int lnet_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
+int lnet_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
+int lnet_sock_getaddr(struct socket *socket, bool remote, __u32 *ip, int *port);
+int lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
+int lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
+
+int lnet_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
+int lnet_sock_accept(struct socket **newsockp, struct socket *sock);
+int lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port);
+
void lnet_get_tunables(void);
int lnet_peers_start_down(void);
int lnet_peer_buffer_credits(lnet_ni_t *ni);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index 3bad441..be4de56 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -2620,7 +2620,7 @@ static kib_dev_t *kiblnd_create_dev(char *ifname)
int up;
int rc;
- rc = libcfs_ipif_query(ifname, &up, &ip, &netmask);
+ rc = lnet_ipif_query(ifname, &up, &ip, &netmask);
if (rc != 0) {
CERROR("Can't query IPoIB interface %s: %d\n",
ifname, rc);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index 7586b7e..cb49744 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -968,7 +968,7 @@ ksocknal_accept(lnet_ni_t *ni, struct socket *sock)
__u32 peer_ip;
int peer_port;
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
LIBCFS_ALLOC(cr, sizeof(*cr));
@@ -1378,15 +1378,15 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
ksocknal_txlist_done(ni, &zombies, 1);
ksocknal_peer_decref(peer);
- failed_1:
+failed_1:
if (hello != NULL)
LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
kshm_ips[LNET_MAX_INTERFACES]));
LIBCFS_FREE(conn, sizeof(*conn));
- failed_0:
- libcfs_sock_release(sock);
+failed_0:
+ sock_release(sock);
return rc;
}
@@ -2594,7 +2594,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
int rc;
int n;
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Can't enumerate interfaces: %d\n", n);
return n;
@@ -2608,7 +2608,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &mask);
+ rc = lnet_ipif_query(names[i], &up, &ip, &mask);
if (rc != 0) {
CWARN("Can't get interface %s info: %d\n",
names[i], rc);
@@ -2634,7 +2634,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
j++;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
if (j == 0)
CERROR("Can't find any usable interfaces\n");
@@ -2796,7 +2796,7 @@ ksocknal_startup(lnet_ni_t *ni)
if (ni->ni_interfaces[i] == NULL)
break;
- rc = libcfs_ipif_query(
+ rc = lnet_ipif_query(
ni->ni_interfaces[i], &up,
&net->ksnn_interfaces[i].ksni_ipaddr,
&net->ksnn_interfaces[i].ksni_netmask);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index c54c995..f3b4923 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -453,7 +453,7 @@ ksocknal_connsock_decref(ksock_conn_t *conn)
LASSERT(atomic_read(&conn->ksnc_sock_refcount) > 0);
if (atomic_dec_and_test(&conn->ksnc_sock_refcount)) {
LASSERT(conn->ksnc_closing);
- libcfs_sock_release(conn->ksnc_sock);
+ sock_release(conn->ksnc_sock);
conn->ksnc_sock = NULL;
ksocknal_finalize_zcreq(conn);
}
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
index fa7ad88..718a8e2 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
@@ -1707,7 +1707,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
timeout = active ? *ksocknal_tunables.ksnd_timeout :
lnet_acceptor_timeout();
- rc = libcfs_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_magic,
+ sizeof(hello->kshm_magic), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -1726,8 +1727,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &hello->kshm_version,
- sizeof(hello->kshm_version), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_version,
+ sizeof(hello->kshm_version), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index f5e8ab0..a93f59c 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -39,9 +39,9 @@
int
ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
{
- int rc = libcfs_sock_getaddr(conn->ksnc_sock, 1,
- &conn->ksnc_ipaddr,
- &conn->ksnc_port);
+ int rc = lnet_sock_getaddr(conn->ksnc_sock, 1,
+ &conn->ksnc_ipaddr,
+ &conn->ksnc_port);
/* Didn't need the {get,put}connsock dance to deref ksnc_sock... */
LASSERT(!conn->ksnc_closing);
@@ -51,8 +51,8 @@ ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
return rc;
}
- rc = libcfs_sock_getaddr(conn->ksnc_sock, 0,
- &conn->ksnc_myipaddr, NULL);
+ rc = lnet_sock_getaddr(conn->ksnc_sock, 0,
+ &conn->ksnc_myipaddr, NULL);
if (rc != 0) {
CERROR("Error %d getting sock local IP\n", rc);
return rc;
@@ -436,7 +436,7 @@ ksocknal_lib_get_conn_tunables(ksock_conn_t *conn, int *txmem, int *rxmem, int *
return -ESHUTDOWN;
}
- rc = libcfs_sock_getbuf(sock, txmem, rxmem);
+ rc = lnet_sock_getbuf(sock, txmem, rxmem);
if (rc == 0) {
len = sizeof(*nagle);
rc = kernel_getsockopt(sock, SOL_TCP, TCP_NODELAY,
@@ -498,9 +498,8 @@ ksocknal_lib_setup_sock(struct socket *sock)
}
}
- rc = libcfs_sock_setbuf(sock,
- *ksocknal_tunables.ksnd_tx_buffer_size,
- *ksocknal_tunables.ksnd_rx_buffer_size);
+ rc = lnet_sock_setbuf(sock, *ksocknal_tunables.ksnd_tx_buffer_size,
+ *ksocknal_tunables.ksnd_rx_buffer_size);
if (rc != 0) {
CERROR("Can't set buffer tx %d, rx %d buffers: %d\n",
*ksocknal_tunables.ksnd_tx_buffer_size,
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
index 8596581..3916790 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
@@ -495,8 +495,8 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hdr->msg.hello.type = cpu_to_le32 (hello->kshm_ctype);
hdr->msg.hello.incarnation = cpu_to_le64 (hello->kshm_src_incarnation);
- rc = libcfs_sock_write(sock, hdr, sizeof(*hdr),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hdr, sizeof(*hdr),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
@@ -511,9 +511,9 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hello->kshm_ips[i] = __cpu_to_le32 (hello->kshm_ips[i]);
}
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -544,8 +544,8 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
LNET_UNLOCK();
}
- rc = libcfs_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
@@ -556,9 +556,9 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -583,7 +583,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
return -ENOMEM;
}
- rc = libcfs_sock_read(sock, &hdr->src_nid,
+ rc = lnet_sock_read(sock, &hdr->src_nid,
sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
timeout);
if (rc != 0) {
@@ -619,7 +619,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
if (hello->kshm_nips == 0)
goto out;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
+ rc = lnet_sock_read(sock, hello->kshm_ips,
hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
@@ -656,7 +656,7 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
else
conn->ksnc_flip = 1;
- rc = libcfs_sock_read(sock, &hello->kshm_src_nid,
+ rc = lnet_sock_read(sock, &hello->kshm_src_nid,
offsetof(ksock_hello_msg_t, kshm_ips) -
offsetof(ksock_hello_msg_t, kshm_src_nid),
timeout);
@@ -687,8 +687,8 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/Makefile b/drivers/staging/lustre/lnet/lnet/Makefile
index 336b8ea..52492fb 100644
--- a/drivers/staging/lustre/lnet/lnet/Makefile
+++ b/drivers/staging/lustre/lnet/lnet/Makefile
@@ -1,5 +1,6 @@
obj-$(CONFIG_LNET) += lnet.o
-lnet-y := api-ni.o config.o lib-me.o lib-msg.o lib-eq.o \
- lib-md.o lib-ptl.o lib-move.o module.o lo.o router.o \
- router_proc.o acceptor.o peer.o
+lnet-y := api-ni.o config.o \
+ lib-me.o lib-msg.o lib-eq.o lib-md.o lib-ptl.o \
+ lib-socket.o lib-move.o module.o lo.o \
+ router.o router_proc.o acceptor.o peer.o
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 72fd1bf..bd97099 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -155,9 +155,8 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
--port) {
/* Iterate through reserved ports. */
- rc = libcfs_sock_connect(&sock, &fatal,
- local_ip, port,
- peer_ip, peer_port);
+ rc = lnet_sock_connect(&sock, &fatal, local_ip, port,
+ peer_ip, peer_port);
if (rc != 0) {
if (fatal)
goto failed;
@@ -184,8 +183,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
lnet_net_unlock(LNET_LOCK_EX);
}
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
goto failed_sock;
@@ -197,7 +195,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
goto failed;
failed_sock:
- libcfs_sock_release(sock);
+ sock_release(sock);
failed:
lnet_connect_console_error(rc, peer_nid, peer_ip, peer_port);
return rc;
@@ -220,7 +218,7 @@ lnet_accept(struct socket *sock, __u32 magic)
LASSERT(sizeof(cr) <= 16); /* not too big for the stack */
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
if (!lnet_accept_magic(magic, LNET_PROTO_ACCEPTOR_MAGIC)) {
@@ -234,7 +232,7 @@ lnet_accept(struct socket *sock, __u32 magic)
memset(&cr, 0, sizeof(cr));
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
accept_timeout);
if (rc != 0)
@@ -257,9 +255,8 @@ lnet_accept(struct socket *sock, __u32 magic)
flip = (magic != LNET_PROTO_ACCEPTOR_MAGIC);
- rc = libcfs_sock_read(sock, &cr.acr_version,
- sizeof(cr.acr_version),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_version, sizeof(cr.acr_version),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request version from %pI4h\n",
rc, &peer_ip);
@@ -280,7 +277,7 @@ lnet_accept(struct socket *sock, __u32 magic)
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
accept_timeout);
if (rc != 0)
@@ -289,7 +286,7 @@ lnet_accept(struct socket *sock, __u32 magic)
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &cr.acr_nid,
+ rc = lnet_sock_read(sock, &cr.acr_nid,
sizeof(cr) -
offsetof(lnet_acceptor_connreq_t, acr_nid),
accept_timeout);
@@ -343,8 +340,8 @@ lnet_acceptor(void *arg)
cfs_block_allsigs();
- rc = libcfs_sock_listen(&lnet_acceptor_state.pta_sock,
- 0, accept_port, accept_backlog);
+ rc = lnet_sock_listen(&lnet_acceptor_state.pta_sock, 0, accept_port,
+ accept_backlog);
if (rc != 0) {
if (rc == -EADDRINUSE)
LCONSOLE_ERROR_MSG(0x122, "Can't start acceptor on port %d: port already in use\n",
@@ -367,7 +364,7 @@ lnet_acceptor(void *arg)
while (!lnet_acceptor_state.pta_shutdown) {
- rc = libcfs_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
+ rc = lnet_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
if (rc != 0) {
if (rc != -EAGAIN) {
CWARN("Accept error %d: pausing...\n", rc);
@@ -377,13 +374,13 @@ lnet_acceptor(void *arg)
continue;
}
- /* maybe we're waken up with libcfs_sock_abort_accept() */
+ /* maybe we're waken up with lnet_sock_abort_accept() */
if (lnet_acceptor_state.pta_shutdown) {
- libcfs_sock_release(newsock);
+ sock_release(newsock);
break;
}
- rc = libcfs_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
if (rc != 0) {
CERROR("Can't determine new connection's address\n");
goto failed;
@@ -395,7 +392,7 @@ lnet_acceptor(void *arg)
goto failed;
}
- rc = libcfs_sock_read(newsock, &magic, sizeof(magic),
+ rc = lnet_sock_read(newsock, &magic, sizeof(magic),
accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
@@ -410,10 +407,10 @@ lnet_acceptor(void *arg)
continue;
failed:
- libcfs_sock_release(newsock);
+ sock_release(newsock);
}
- libcfs_sock_release(lnet_acceptor_state.pta_sock);
+ sock_release(lnet_acceptor_state.pta_sock);
lnet_acceptor_state.pta_sock = NULL;
CDEBUG(D_NET, "Acceptor stopping\n");
@@ -493,7 +490,7 @@ lnet_acceptor_stop(void)
return;
lnet_acceptor_state.pta_shutdown = 1;
- libcfs_sock_abort_accept(lnet_acceptor_state.pta_sock);
+ wake_up_all(sk_sleep(lnet_acceptor_state.pta_sock->sk));
/* block until acceptor signals exit */
wait_for_completion(&lnet_acceptor_state.pta_signal);
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index 2dc4c4a..d8c8f2e 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -1118,7 +1118,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
__u32 *ipaddrs2;
int nip;
char **ifnames;
- int nif = libcfs_ipif_enumerate(&ifnames);
+ int nif = lnet_ipif_enumerate(&ifnames);
int i;
int rc;
@@ -1128,7 +1128,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
LIBCFS_ALLOC(ipaddrs, nif * sizeof(*ipaddrs));
if (ipaddrs == NULL) {
CERROR("Can't allocate ipaddrs[%d]\n", nif);
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
return -ENOMEM;
}
@@ -1136,8 +1136,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
if (!strcmp(ifnames[i], "lo"))
continue;
- rc = libcfs_ipif_query(ifnames[i], &up,
- &ipaddrs[nip], &netmask);
+ rc = lnet_ipif_query(ifnames[i], &up, &ipaddrs[nip], &netmask);
if (rc != 0) {
CWARN("Can't query interface %s: %d\n",
ifnames[i], rc);
@@ -1153,7 +1152,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
nip++;
}
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
if (nip == nif) {
*ipaddrsp = ipaddrs;
@@ -1237,8 +1236,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return -EPERM;
}
- rc = libcfs_ipif_query(ni->ni_interfaces[0],
- &up, &ip, &netmask);
+ rc = lnet_ipif_query(ni->ni_interfaces[0], &up, &ip, &netmask);
if (rc != 0) {
CERROR("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), ni->ni_interfaces[0], rc);
@@ -1255,7 +1253,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return 0;
}
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Net %s can't enumerate interfaces: %d\n",
libcfs_net2str(net), n);
@@ -1266,8 +1264,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &netmask);
-
+ rc = lnet_ipif_query(names[i], &up, &ip, &netmask);
if (rc != 0) {
CWARN("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), names[i], rc);
@@ -1280,13 +1277,13 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
continue;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
ni->ni_nid = LNET_MKNID(net, ip);
return 0;
}
CERROR("Net %s can't find any interfaces\n", libcfs_net2str(net));
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
return -ENOENT;
}
EXPORT_SYMBOL(lnet_set_ip_niaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
new file mode 100644
index 0000000..da7b871
--- /dev/null
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -0,0 +1,616 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, 2014, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+#define DEBUG_SUBSYSTEM S_LNET
+
+#include <linux/if.h>
+#include <linux/in.h>
+#include <linux/net.h>
+#include <linux/file.h>
+#include <linux/pagemap.h>
+/* For sys_open & sys_close */
+#include <linux/syscalls.h>
+#include <net/sock.h>
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/lib-lnet.h"
+
+static int
+kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
+{
+ mm_segment_t oldfs = get_fs();
+ int err;
+
+ set_fs(KERNEL_DS);
+ err = filp->f_op->unlocked_ioctl(filp, cmd, arg);
+ set_fs(oldfs);
+
+ return err;
+}
+
+static int
+lnet_sock_ioctl(int cmd, unsigned long arg)
+{
+ struct file *sock_filp;
+ struct socket *sock;
+ int fd = -1;
+ int rc;
+
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
+ if (rc != 0) {
+ CERROR("Can't create socket: %d\n", rc);
+ return rc;
+ }
+
+ sock_filp = sock_alloc_file(sock, 0, NULL);
+ if (!sock_filp) {
+ rc = -ENOMEM;
+ sock_release(sock);
+ goto out;
+ }
+
+ rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
+
+ fput(sock_filp);
+out:
+ if (fd >= 0)
+ sys_close(fd);
+ return rc;
+}
+
+int
+lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask)
+{
+ struct ifreq ifr;
+ int nob;
+ int rc;
+ __u32 val;
+
+ nob = strnlen(name, IFNAMSIZ);
+ if (nob == IFNAMSIZ) {
+ CERROR("Interface name %s too long\n", name);
+ return -EINVAL;
+ }
+
+ CLASSERT(sizeof(ifr.ifr_name) >= IFNAMSIZ);
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
+ if (rc != 0) {
+ CERROR("Can't get flags for interface %s\n", name);
+ return rc;
+ }
+
+ if ((ifr.ifr_flags & IFF_UP) == 0) {
+ CDEBUG(D_NET, "Interface %s down\n", name);
+ *up = 0;
+ *ip = *mask = 0;
+ return 0;
+ }
+ *up = 1;
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ ifr.ifr_addr.sa_family = AF_INET;
+ rc = lnet_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
+
+ if (rc != 0) {
+ CERROR("Can't get IP address for interface %s\n", name);
+ return rc;
+ }
+
+ val = ((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr;
+ *ip = ntohl(val);
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ ifr.ifr_addr.sa_family = AF_INET;
+ rc = lnet_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
+ if (rc != 0) {
+ CERROR("Can't get netmask for interface %s\n", name);
+ return rc;
+ }
+
+ val = ((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr.s_addr;
+ *mask = ntohl(val);
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_ipif_query);
+
+void
+lnet_ipif_free_enumeration(char **names, int n)
+{
+ int i;
+
+ LASSERT(n > 0);
+
+ for (i = 0; i < n && names[i] != NULL; i++)
+ LIBCFS_FREE(names[i], IFNAMSIZ);
+
+ LIBCFS_FREE(names, n * sizeof(*names));
+}
+EXPORT_SYMBOL(lnet_ipif_free_enumeration);
+
+int
+lnet_ipif_enumerate(char ***namesp)
+{
+ /* Allocate and fill in 'names', returning # interfaces/error */
+ char **names;
+ int toobig;
+ int nalloc;
+ int nfound;
+ struct ifreq *ifr;
+ struct ifconf ifc;
+ int rc;
+ int nob;
+ int i;
+
+ nalloc = 16; /* first guess@max interfaces */
+ toobig = 0;
+ for (;;) {
+ if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
+ toobig = 1;
+ nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
+ CWARN("Too many interfaces: only enumerating "
+ "first %d\n", nalloc);
+ }
+
+ LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
+ if (ifr == NULL) {
+ CERROR("ENOMEM enumerating up to %d interfaces\n",
+ nalloc);
+ rc = -ENOMEM;
+ goto out0;
+ }
+
+ ifc.ifc_buf = (char *)ifr;
+ ifc.ifc_len = nalloc * sizeof(*ifr);
+
+ rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
+ if (rc < 0) {
+ CERROR("Error %d enumerating interfaces\n", rc);
+ goto out1;
+ }
+
+ LASSERT(rc == 0);
+
+ nfound = ifc.ifc_len/sizeof(*ifr);
+ LASSERT(nfound <= nalloc);
+
+ if (nfound < nalloc || toobig)
+ break;
+
+ LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
+ nalloc *= 2;
+ }
+
+ if (nfound == 0)
+ goto out1;
+
+ LIBCFS_ALLOC(names, nfound * sizeof(*names));
+ if (names == NULL) {
+ rc = -ENOMEM;
+ goto out1;
+ }
+
+ for (i = 0; i < nfound; i++) {
+ nob = strnlen(ifr[i].ifr_name, IFNAMSIZ);
+ if (nob == IFNAMSIZ) {
+ /* no space for terminating NULL */
+ CERROR("interface name %.*s too long (%d max)\n",
+ nob, ifr[i].ifr_name, IFNAMSIZ);
+ rc = -ENAMETOOLONG;
+ goto out2;
+ }
+
+ LIBCFS_ALLOC(names[i], IFNAMSIZ);
+ if (names[i] == NULL) {
+ rc = -ENOMEM;
+ goto out2;
+ }
+
+ memcpy(names[i], ifr[i].ifr_name, nob);
+ names[i][nob] = 0;
+ }
+
+ *namesp = names;
+ rc = nfound;
+
+ out2:
+ if (rc < 0)
+ lnet_ipif_free_enumeration(names, nfound);
+ out1:
+ LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
+ out0:
+ return rc;
+}
+EXPORT_SYMBOL(lnet_ipif_enumerate);
+
+int
+lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout)
+{
+ int rc;
+ long jiffies_left = timeout * msecs_to_jiffies(MSEC_PER_SEC);
+ unsigned long then;
+ struct timeval tv;
+
+ LASSERT(nob > 0);
+ /* Caller may pass a zero timeout if she thinks the socket buffer is
+ * empty enough to take the whole message immediately */
+
+ for (;;) {
+ struct kvec iov = {
+ .iov_base = buffer,
+ .iov_len = nob
+ };
+ struct msghdr msg = {
+ .msg_flags = (timeout == 0) ? MSG_DONTWAIT : 0
+ };
+
+ if (timeout != 0) {
+ /* Set send timeout to remaining time */
+ tv = (struct timeval) {
+ .tv_sec = jiffies_left /
+ msecs_to_jiffies(MSEC_PER_SEC),
+ .tv_usec = ((jiffies_left %
+ msecs_to_jiffies(MSEC_PER_SEC)) *
+ USEC_PER_SEC) /
+ msecs_to_jiffies(MSEC_PER_SEC)
+ };
+
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
+ (char *)&tv, sizeof(tv));
+ if (rc != 0) {
+ CERROR("Can't set socket send timeout "
+ "%ld.%06d: %d\n",
+ (long)tv.tv_sec, (int)tv.tv_usec, rc);
+ return rc;
+ }
+ }
+
+ then = jiffies;
+ rc = kernel_sendmsg(sock, &msg, &iov, 1, nob);
+ jiffies_left -= jiffies - then;
+
+ if (rc == nob)
+ return 0;
+
+ if (rc < 0)
+ return rc;
+
+ if (rc == 0) {
+ CERROR("Unexpected zero rc\n");
+ return -ECONNABORTED;
+ }
+
+ if (jiffies_left <= 0)
+ return -EAGAIN;
+
+ buffer = ((char *)buffer) + rc;
+ nob -= rc;
+ }
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_write);
+
+int
+lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout)
+{
+ int rc;
+ long jiffies_left = timeout * msecs_to_jiffies(MSEC_PER_SEC);
+ unsigned long then;
+ struct timeval tv;
+
+ LASSERT(nob > 0);
+ LASSERT(jiffies_left > 0);
+
+ for (;;) {
+ struct kvec iov = {
+ .iov_base = buffer,
+ .iov_len = nob
+ };
+ struct msghdr msg = {
+ .msg_flags = 0
+ };
+
+ /* Set receive timeout to remaining time */
+ tv = (struct timeval) {
+ .tv_sec = jiffies_left / msecs_to_jiffies(MSEC_PER_SEC),
+ .tv_usec = ((jiffies_left %
+ msecs_to_jiffies(MSEC_PER_SEC)) *
+ USEC_PER_SEC) /
+ msecs_to_jiffies(MSEC_PER_SEC)
+ };
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
+ (char *)&tv, sizeof(tv));
+ if (rc != 0) {
+ CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
+ (long)tv.tv_sec, (int)tv.tv_usec, rc);
+ return rc;
+ }
+
+ then = jiffies;
+ rc = kernel_recvmsg(sock, &msg, &iov, 1, nob, 0);
+ jiffies_left -= jiffies - then;
+
+ if (rc < 0)
+ return rc;
+
+ if (rc == 0)
+ return -ECONNRESET;
+
+ buffer = ((char *)buffer) + rc;
+ nob -= rc;
+
+ if (nob == 0)
+ return 0;
+
+ if (jiffies_left <= 0)
+ return -ETIMEDOUT;
+ }
+}
+EXPORT_SYMBOL(lnet_sock_read);
+
+static int
+lnet_sock_create(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port)
+{
+ struct sockaddr_in locaddr;
+ struct socket *sock;
+ int rc;
+ int option;
+
+ /* All errors are fatal except bind failure if the port is in use */
+ *fatal = 1;
+
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
+ *sockp = sock;
+ if (rc != 0) {
+ CERROR("Can't create socket: %d\n", rc);
+ return rc;
+ }
+
+ option = 1;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
+ goto failed;
+ }
+
+ if (local_ip != 0 || local_port != 0) {
+ memset(&locaddr, 0, sizeof(locaddr));
+ locaddr.sin_family = AF_INET;
+ locaddr.sin_port = htons(local_port);
+ locaddr.sin_addr.s_addr = (local_ip == 0) ?
+ INADDR_ANY : htonl(local_ip);
+
+ rc = kernel_bind(sock, (struct sockaddr *)&locaddr,
+ sizeof(locaddr));
+ if (rc == -EADDRINUSE) {
+ CDEBUG(D_NET, "Port %d already in use\n", local_port);
+ *fatal = 0;
+ goto failed;
+ }
+ if (rc != 0) {
+ CERROR("Error trying to bind to port %d: %d\n",
+ local_port, rc);
+ goto failed;
+ }
+ }
+ return 0;
+
+failed:
+ sock_release(sock);
+ return rc;
+}
+
+int
+lnet_sock_setbuf(struct socket *sock, int txbufsize, int rxbufsize)
+{
+ int option;
+ int rc;
+
+ if (txbufsize != 0) {
+ option = txbufsize;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set send buffer %d: %d\n",
+ option, rc);
+ return rc;
+ }
+ }
+
+ if (rxbufsize != 0) {
+ option = rxbufsize;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set receive buffer %d: %d\n",
+ option, rc);
+ return rc;
+ }
+ }
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_setbuf);
+
+int
+lnet_sock_getaddr(struct socket *sock, bool remote, __u32 *ip, int *port)
+{
+ struct sockaddr_in sin;
+ int len = sizeof(sin);
+ int rc;
+
+ if (remote)
+ rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
+ else
+ rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
+ if (rc != 0) {
+ CERROR("Error %d getting sock %s IP/port\n",
+ rc, remote ? "peer" : "local");
+ return rc;
+ }
+
+ if (ip != NULL)
+ *ip = ntohl(sin.sin_addr.s_addr);
+
+ if (port != NULL)
+ *port = ntohs(sin.sin_port);
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_getaddr);
+
+int
+lnet_sock_getbuf(struct socket *sock, int *txbufsize, int *rxbufsize)
+{
+ if (txbufsize != NULL)
+ *txbufsize = sock->sk->sk_sndbuf;
+
+ if (rxbufsize != NULL)
+ *rxbufsize = sock->sk->sk_rcvbuf;
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_getbuf);
+
+int
+lnet_sock_listen(struct socket **sockp,
+ __u32 local_ip, int local_port, int backlog)
+{
+ int fatal;
+ int rc;
+
+ rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
+ if (rc != 0) {
+ if (!fatal)
+ CERROR("Can't create socket: port %d already in use\n",
+ local_port);
+ return rc;
+ }
+
+ rc = kernel_listen(*sockp, backlog);
+ if (rc == 0)
+ return 0;
+
+ CERROR("Can't set listen backlog %d: %d\n", backlog, rc);
+ sock_release(*sockp);
+ return rc;
+}
+
+int
+lnet_sock_accept(struct socket **newsockp, struct socket *sock)
+{
+ wait_queue_t wait;
+ struct socket *newsock;
+ int rc;
+
+ /* XXX this should add a ref to sock->ops->owner, if
+ * TCP could be a module */
+ rc = sock_create_lite(PF_PACKET, sock->type, IPPROTO_TCP, &newsock);
+ if (rc) {
+ CERROR("Can't allocate socket\n");
+ return rc;
+ }
+
+ newsock->ops = sock->ops;
+
+ rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
+ if (rc == -EAGAIN) {
+ /* Nothing ready, so wait for activity */
+ init_waitqueue_entry(&wait, current);
+ add_wait_queue(sk_sleep(sock->sk), &wait);
+ set_current_state(TASK_INTERRUPTIBLE);
+ schedule();
+ remove_wait_queue(sk_sleep(sock->sk), &wait);
+ rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
+ }
+
+ if (rc != 0)
+ goto failed;
+
+ *newsockp = newsock;
+ return 0;
+
+failed:
+ sock_release(newsock);
+ return rc;
+}
+
+int
+lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port)
+{
+ struct sockaddr_in srvaddr;
+ int rc;
+
+ rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
+ if (rc != 0)
+ return rc;
+
+ memset(&srvaddr, 0, sizeof(srvaddr));
+ srvaddr.sin_family = AF_INET;
+ srvaddr.sin_port = htons(peer_port);
+ srvaddr.sin_addr.s_addr = htonl(peer_ip);
+
+ rc = kernel_connect(*sockp, (struct sockaddr *)&srvaddr,
+ sizeof(srvaddr), 0);
+ if (rc == 0)
+ return 0;
+
+ /* EADDRNOTAVAIL probably means we're already connected to the same
+ * peer/port on the same local port on a differently typed
+ * connection. Let our caller retry with a different local
+ * port... */
+ *fatal = !(rc == -EADDRNOTAVAIL);
+
+ CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
+ "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
+ &local_ip, local_port, &peer_ip, peer_port);
+
+ sock_release(*sockp);
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/libcfs/Makefile b/drivers/staging/lustre/lustre/libcfs/Makefile
index fabdd3e..ec98f44 100644
--- a/drivers/staging/lustre/lustre/libcfs/Makefile
+++ b/drivers/staging/lustre/lustre/libcfs/Makefile
@@ -2,7 +2,6 @@ obj-$(CONFIG_LUSTRE_FS) += libcfs.o
libcfs-linux-objs := linux-tracefile.o linux-debug.o
libcfs-linux-objs += linux-prim.o linux-cpu.o
-libcfs-linux-objs += linux-tcpip.o
libcfs-linux-objs += linux-curproc.o
libcfs-linux-objs += linux-module.o
libcfs-linux-objs += linux-crypto.o
diff --git a/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c b/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
deleted file mode 100644
index f2462e7..0000000
--- a/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
+++ /dev/null
@@ -1,623 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-#define DEBUG_SUBSYSTEM S_LNET
-
-#include "../../../include/linux/libcfs/libcfs.h"
-
-#include <linux/if.h>
-#include <linux/in.h>
-#include <linux/file.h>
-/* For sys_open & sys_close */
-#include <linux/syscalls.h>
-
-static int
-libcfs_sock_ioctl(int cmd, unsigned long arg)
-{
- mm_segment_t oldmm = get_fs();
- struct socket *sock;
- int rc;
- struct file *sock_filp;
-
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
- if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
- return rc;
- }
-
- sock_filp = sock_alloc_file(sock, 0, NULL);
- if (IS_ERR(sock_filp)) {
- sock_release(sock);
- rc = PTR_ERR(sock_filp);
- goto out;
- }
-
- set_fs(KERNEL_DS);
- if (sock_filp->f_op->unlocked_ioctl)
- rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
- set_fs(oldmm);
-
- fput(sock_filp);
-out:
- return rc;
-}
-
-int
-libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
-{
- struct ifreq ifr;
- int nob;
- int rc;
- __u32 val;
-
- nob = strnlen(name, IFNAMSIZ);
- if (nob == IFNAMSIZ) {
- CERROR("Interface name %s too long\n", name);
- return -EINVAL;
- }
-
- CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
-
- strcpy(ifr.ifr_name, name);
- rc = libcfs_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get flags for interface %s\n", name);
- return rc;
- }
-
- if ((ifr.ifr_flags & IFF_UP) == 0) {
- CDEBUG(D_NET, "Interface %s down\n", name);
- *up = 0;
- *ip = *mask = 0;
- return 0;
- }
-
- *up = 1;
-
- strcpy(ifr.ifr_name, name);
- ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get IP address for interface %s\n", name);
- return rc;
- }
-
- val = ((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr;
- *ip = ntohl(val);
-
- strcpy(ifr.ifr_name, name);
- ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get netmask for interface %s\n", name);
- return rc;
- }
-
- val = ((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr.s_addr;
- *mask = ntohl(val);
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_ipif_query);
-
-int
-libcfs_ipif_enumerate (char ***namesp)
-{
- /* Allocate and fill in 'names', returning # interfaces/error */
- char **names;
- int toobig;
- int nalloc;
- int nfound;
- struct ifreq *ifr;
- struct ifconf ifc;
- int rc;
- int nob;
- int i;
-
-
- nalloc = 16; /* first guess at max interfaces */
- toobig = 0;
- for (;;) {
- if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
- toobig = 1;
- nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
- CWARN("Too many interfaces: only enumerating first %d\n",
- nalloc);
- }
-
- LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
- if (ifr == NULL) {
- CERROR ("ENOMEM enumerating up to %d interfaces\n", nalloc);
- rc = -ENOMEM;
- goto out0;
- }
-
- ifc.ifc_buf = (char *)ifr;
- ifc.ifc_len = nalloc * sizeof(*ifr);
-
- rc = libcfs_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
-
- if (rc < 0) {
- CERROR ("Error %d enumerating interfaces\n", rc);
- goto out1;
- }
-
- LASSERT (rc == 0);
-
- nfound = ifc.ifc_len/sizeof(*ifr);
- LASSERT (nfound <= nalloc);
-
- if (nfound < nalloc || toobig)
- break;
-
- LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- nalloc *= 2;
- }
-
- if (nfound == 0)
- goto out1;
-
- LIBCFS_ALLOC(names, nfound * sizeof(*names));
- if (names == NULL) {
- rc = -ENOMEM;
- goto out1;
- }
-
- for (i = 0; i < nfound; i++) {
-
- nob = strnlen (ifr[i].ifr_name, IFNAMSIZ);
- if (nob == IFNAMSIZ) {
- /* no space for terminating NULL */
- CERROR("interface name %.*s too long (%d max)\n",
- nob, ifr[i].ifr_name, IFNAMSIZ);
- rc = -ENAMETOOLONG;
- goto out2;
- }
-
- LIBCFS_ALLOC(names[i], IFNAMSIZ);
- if (names[i] == NULL) {
- rc = -ENOMEM;
- goto out2;
- }
-
- memcpy(names[i], ifr[i].ifr_name, nob);
- names[i][nob] = 0;
- }
-
- *namesp = names;
- rc = nfound;
-
- out2:
- if (rc < 0)
- libcfs_ipif_free_enumeration(names, nfound);
- out1:
- LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- out0:
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_ipif_enumerate);
-
-void
-libcfs_ipif_free_enumeration (char **names, int n)
-{
- int i;
-
- LASSERT (n > 0);
-
- for (i = 0; i < n && names[i] != NULL; i++)
- LIBCFS_FREE(names[i], IFNAMSIZ);
-
- LIBCFS_FREE(names, n * sizeof(*names));
-}
-
-EXPORT_SYMBOL(libcfs_ipif_free_enumeration);
-
-int
-libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
-{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
- struct timeval tv;
-
- LASSERT (nob > 0);
- /* Caller may pass a zero timeout if she thinks the socket buffer is
- * empty enough to take the whole message immediately */
-
- for (;;) {
- struct kvec iov = {
- .iov_base = buffer,
- .iov_len = nob
- };
- struct msghdr msg = {
- .msg_flags = (timeout == 0) ? MSG_DONTWAIT : 0
- };
-
- if (timeout != 0) {
- /* Set send timeout to remaining time */
- tv = (struct timeval) {
- .tv_sec = ticks / HZ,
- .tv_usec = ((ticks % HZ) * 1000000) / HZ
- };
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
- (char *)&tv, sizeof(tv));
- if (rc != 0) {
- CERROR("Can't set socket send timeout %ld.%06d: %d\n",
- (long)tv.tv_sec, (int)tv.tv_usec, rc);
- return rc;
- }
- }
-
- then = jiffies;
- rc = kernel_sendmsg(sock, &msg, &iov, 1, nob);
- ticks -= jiffies - then;
-
- if (rc == nob)
- return 0;
-
- if (rc < 0)
- return rc;
-
- if (rc == 0) {
- CERROR ("Unexpected zero rc\n");
- return -ECONNABORTED;
- }
-
- if (ticks <= 0)
- return -EAGAIN;
-
- buffer = ((char *)buffer) + rc;
- nob -= rc;
- }
-
- return 0;
-}
-EXPORT_SYMBOL(libcfs_sock_write);
-
-int
-libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
-{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
- struct timeval tv;
-
- LASSERT (nob > 0);
- LASSERT (ticks > 0);
-
- for (;;) {
- struct kvec iov = {
- .iov_base = buffer,
- .iov_len = nob
- };
- struct msghdr msg = {
- .msg_flags = 0
- };
-
- /* Set receive timeout to remaining time */
- tv = (struct timeval) {
- .tv_sec = ticks / HZ,
- .tv_usec = ((ticks % HZ) * 1000000) / HZ
- };
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
- (char *)&tv, sizeof(tv));
- if (rc != 0) {
- CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
- (long)tv.tv_sec, (int)tv.tv_usec, rc);
- return rc;
- }
-
- then = jiffies;
- rc = kernel_recvmsg(sock, &msg, &iov, 1, nob, 0);
- ticks -= jiffies - then;
-
- if (rc < 0)
- return rc;
-
- if (rc == 0)
- return -ECONNRESET;
-
- buffer = ((char *)buffer) + rc;
- nob -= rc;
-
- if (nob == 0)
- return 0;
-
- if (ticks <= 0)
- return -ETIMEDOUT;
- }
-}
-
-EXPORT_SYMBOL(libcfs_sock_read);
-
-static int
-libcfs_sock_create (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port)
-{
- struct sockaddr_in locaddr;
- struct socket *sock;
- int rc;
- int option;
-
- /* All errors are fatal except bind failure if the port is in use */
- *fatal = 1;
-
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
- *sockp = sock;
- if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
- return rc;
- }
-
- option = 1;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
- goto failed;
- }
-
- if (local_ip != 0 || local_port != 0) {
- memset(&locaddr, 0, sizeof(locaddr));
- locaddr.sin_family = AF_INET;
- locaddr.sin_port = htons(local_port);
- locaddr.sin_addr.s_addr = (local_ip == 0) ?
- INADDR_ANY : htonl(local_ip);
-
- rc = sock->ops->bind(sock, (struct sockaddr *)&locaddr,
- sizeof(locaddr));
- if (rc == -EADDRINUSE) {
- CDEBUG(D_NET, "Port %d already in use\n", local_port);
- *fatal = 0;
- goto failed;
- }
- if (rc != 0) {
- CERROR("Error trying to bind to port %d: %d\n",
- local_port, rc);
- goto failed;
- }
- }
-
- return 0;
-
- failed:
- sock_release(sock);
- return rc;
-}
-
-int
-libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
-{
- int option;
- int rc;
-
- if (txbufsize != 0) {
- option = txbufsize;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR ("Can't set send buffer %d: %d\n",
- option, rc);
- return rc;
- }
- }
-
- if (rxbufsize != 0) {
- option = rxbufsize;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR ("Can't set receive buffer %d: %d\n",
- option, rc);
- return rc;
- }
- }
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_setbuf);
-
-int
-libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
-{
- struct sockaddr_in sin;
- int len = sizeof (sin);
- int rc;
-
- rc = sock->ops->getname (sock, (struct sockaddr *)&sin, &len,
- remote ? 2 : 0);
- if (rc != 0) {
- CERROR ("Error %d getting sock %s IP/port\n",
- rc, remote ? "peer" : "local");
- return rc;
- }
-
- if (ip != NULL)
- *ip = ntohl (sin.sin_addr.s_addr);
-
- if (port != NULL)
- *port = ntohs (sin.sin_port);
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_getaddr);
-
-int
-libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
-{
-
- if (txbufsize != NULL) {
- *txbufsize = sock->sk->sk_sndbuf;
- }
-
- if (rxbufsize != NULL) {
- *rxbufsize = sock->sk->sk_rcvbuf;
- }
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_getbuf);
-
-int
-libcfs_sock_listen (struct socket **sockp,
- __u32 local_ip, int local_port, int backlog)
-{
- int fatal;
- int rc;
-
- rc = libcfs_sock_create(sockp, &fatal, local_ip, local_port);
- if (rc != 0) {
- if (!fatal)
- CERROR("Can't create socket: port %d already in use\n",
- local_port);
- return rc;
- }
-
- rc = (*sockp)->ops->listen(*sockp, backlog);
- if (rc == 0)
- return 0;
-
- CERROR("Can't set listen backlog %d: %d\n", backlog, rc);
- sock_release(*sockp);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_listen);
-
-int
-libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
-{
- wait_queue_t wait;
- struct socket *newsock;
- int rc;
-
- init_waitqueue_entry(&wait, current);
-
- /* XXX this should add a ref to sock->ops->owner, if
- * TCP could be a module */
- rc = sock_create_lite(PF_PACKET, sock->type, IPPROTO_TCP, &newsock);
- if (rc) {
- CERROR("Can't allocate socket\n");
- return rc;
- }
-
- newsock->ops = sock->ops;
-
- rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
- if (rc == -EAGAIN) {
- /* Nothing ready, so wait for activity */
- set_current_state(TASK_INTERRUPTIBLE);
- add_wait_queue(sk_sleep(sock->sk), &wait);
- schedule();
- remove_wait_queue(sk_sleep(sock->sk), &wait);
- set_current_state(TASK_RUNNING);
- rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
- }
-
- if (rc != 0)
- goto failed;
-
- *newsockp = newsock;
- return 0;
-
- failed:
- sock_release(newsock);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_accept);
-
-void
-libcfs_sock_abort_accept (struct socket *sock)
-{
- wake_up_all(sk_sleep(sock->sk));
-}
-
-EXPORT_SYMBOL(libcfs_sock_abort_accept);
-
-int
-libcfs_sock_connect (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port)
-{
- struct sockaddr_in srvaddr;
- int rc;
-
- rc = libcfs_sock_create(sockp, fatal, local_ip, local_port);
- if (rc != 0)
- return rc;
-
- memset (&srvaddr, 0, sizeof (srvaddr));
- srvaddr.sin_family = AF_INET;
- srvaddr.sin_port = htons(peer_port);
- srvaddr.sin_addr.s_addr = htonl(peer_ip);
-
- rc = (*sockp)->ops->connect(*sockp,
- (struct sockaddr *)&srvaddr, sizeof(srvaddr),
- 0);
- if (rc == 0)
- return 0;
-
- /* EADDRNOTAVAIL probably means we're already connected to the same
- * peer/port on the same local port on a differently typed
- * connection. Let our caller retry with a different local
- * port... */
- *fatal = !(rc == -EADDRNOTAVAIL);
-
- CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
- "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
- &local_ip, local_port, &peer_ip, peer_port);
-
- sock_release(*sockp);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_connect);
-
-void
-libcfs_sock_release (struct socket *sock)
-{
- sock_release(sock);
-}
-
-EXPORT_SYMBOL(libcfs_sock_release);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons, James Simmons
From: James Simmons <uja.ornl@yahoo.com>
Since libcfs no longer builds for user land we can
move the TCPIP abstraction that exist to the LNET
layer which is the only place that uses it. Also
the migrated code will use native linux kernel
APIs directly instead of with wrappers.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13760
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 17 -
.../staging/lustre/include/linux/lnet/lib-lnet.h | 15 +
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 16 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 7 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 17 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 30 +-
drivers/staging/lustre/lnet/lnet/Makefile | 7 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 43 +-
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 616 +++++++++++++++++++
drivers/staging/lustre/lustre/libcfs/Makefile | 1 -
.../lustre/lustre/libcfs/linux/linux-tcpip.c | 623 --------------------
14 files changed, 701 insertions(+), 716 deletions(-)
create mode 100644 drivers/staging/lustre/lnet/lnet/lib-socket.c
delete mode 100644 drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 947df7e..f54bab6 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -88,23 +88,6 @@ static inline int __is_po2(unsigned long long val)
int libcfs_arch_init(void);
void libcfs_arch_cleanup(void);
-/* libcfs tcpip */
-int libcfs_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
-int libcfs_ipif_enumerate(char ***names);
-void libcfs_ipif_free_enumeration(char **names, int n);
-int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
-int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-void libcfs_sock_abort_accept(struct socket *sock);
-int libcfs_sock_connect(struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port);
-int libcfs_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
-int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
-int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
-int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
-int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-void libcfs_sock_release(struct socket *sock);
-
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
#define LNET_ACCEPTOR_MAX_RESERVED_PORT 1023
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index 0038d29..c00a2ca 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -846,6 +846,21 @@ int lnet_acceptor_port(void);
int lnet_acceptor_start(void);
void lnet_acceptor_stop(void);
+int lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
+int lnet_ipif_enumerate(char ***names);
+void lnet_ipif_free_enumeration(char **names, int n);
+int lnet_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
+int lnet_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
+int lnet_sock_getaddr(struct socket *socket, bool remote, __u32 *ip, int *port);
+int lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
+int lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
+
+int lnet_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
+int lnet_sock_accept(struct socket **newsockp, struct socket *sock);
+int lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port);
+
void lnet_get_tunables(void);
int lnet_peers_start_down(void);
int lnet_peer_buffer_credits(lnet_ni_t *ni);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index 3bad441..be4de56 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -2620,7 +2620,7 @@ static kib_dev_t *kiblnd_create_dev(char *ifname)
int up;
int rc;
- rc = libcfs_ipif_query(ifname, &up, &ip, &netmask);
+ rc = lnet_ipif_query(ifname, &up, &ip, &netmask);
if (rc != 0) {
CERROR("Can't query IPoIB interface %s: %d\n",
ifname, rc);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index 7586b7e..cb49744 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -968,7 +968,7 @@ ksocknal_accept(lnet_ni_t *ni, struct socket *sock)
__u32 peer_ip;
int peer_port;
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
LIBCFS_ALLOC(cr, sizeof(*cr));
@@ -1378,15 +1378,15 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
ksocknal_txlist_done(ni, &zombies, 1);
ksocknal_peer_decref(peer);
- failed_1:
+failed_1:
if (hello != NULL)
LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
kshm_ips[LNET_MAX_INTERFACES]));
LIBCFS_FREE(conn, sizeof(*conn));
- failed_0:
- libcfs_sock_release(sock);
+failed_0:
+ sock_release(sock);
return rc;
}
@@ -2594,7 +2594,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
int rc;
int n;
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Can't enumerate interfaces: %d\n", n);
return n;
@@ -2608,7 +2608,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &mask);
+ rc = lnet_ipif_query(names[i], &up, &ip, &mask);
if (rc != 0) {
CWARN("Can't get interface %s info: %d\n",
names[i], rc);
@@ -2634,7 +2634,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
j++;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
if (j == 0)
CERROR("Can't find any usable interfaces\n");
@@ -2796,7 +2796,7 @@ ksocknal_startup(lnet_ni_t *ni)
if (ni->ni_interfaces[i] == NULL)
break;
- rc = libcfs_ipif_query(
+ rc = lnet_ipif_query(
ni->ni_interfaces[i], &up,
&net->ksnn_interfaces[i].ksni_ipaddr,
&net->ksnn_interfaces[i].ksni_netmask);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index c54c995..f3b4923 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -453,7 +453,7 @@ ksocknal_connsock_decref(ksock_conn_t *conn)
LASSERT(atomic_read(&conn->ksnc_sock_refcount) > 0);
if (atomic_dec_and_test(&conn->ksnc_sock_refcount)) {
LASSERT(conn->ksnc_closing);
- libcfs_sock_release(conn->ksnc_sock);
+ sock_release(conn->ksnc_sock);
conn->ksnc_sock = NULL;
ksocknal_finalize_zcreq(conn);
}
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
index fa7ad88..718a8e2 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
@@ -1707,7 +1707,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
timeout = active ? *ksocknal_tunables.ksnd_timeout :
lnet_acceptor_timeout();
- rc = libcfs_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_magic,
+ sizeof(hello->kshm_magic), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -1726,8 +1727,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &hello->kshm_version,
- sizeof(hello->kshm_version), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_version,
+ sizeof(hello->kshm_version), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index f5e8ab0..a93f59c 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -39,9 +39,9 @@
int
ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
{
- int rc = libcfs_sock_getaddr(conn->ksnc_sock, 1,
- &conn->ksnc_ipaddr,
- &conn->ksnc_port);
+ int rc = lnet_sock_getaddr(conn->ksnc_sock, 1,
+ &conn->ksnc_ipaddr,
+ &conn->ksnc_port);
/* Didn't need the {get,put}connsock dance to deref ksnc_sock... */
LASSERT(!conn->ksnc_closing);
@@ -51,8 +51,8 @@ ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
return rc;
}
- rc = libcfs_sock_getaddr(conn->ksnc_sock, 0,
- &conn->ksnc_myipaddr, NULL);
+ rc = lnet_sock_getaddr(conn->ksnc_sock, 0,
+ &conn->ksnc_myipaddr, NULL);
if (rc != 0) {
CERROR("Error %d getting sock local IP\n", rc);
return rc;
@@ -436,7 +436,7 @@ ksocknal_lib_get_conn_tunables(ksock_conn_t *conn, int *txmem, int *rxmem, int *
return -ESHUTDOWN;
}
- rc = libcfs_sock_getbuf(sock, txmem, rxmem);
+ rc = lnet_sock_getbuf(sock, txmem, rxmem);
if (rc == 0) {
len = sizeof(*nagle);
rc = kernel_getsockopt(sock, SOL_TCP, TCP_NODELAY,
@@ -498,9 +498,8 @@ ksocknal_lib_setup_sock(struct socket *sock)
}
}
- rc = libcfs_sock_setbuf(sock,
- *ksocknal_tunables.ksnd_tx_buffer_size,
- *ksocknal_tunables.ksnd_rx_buffer_size);
+ rc = lnet_sock_setbuf(sock, *ksocknal_tunables.ksnd_tx_buffer_size,
+ *ksocknal_tunables.ksnd_rx_buffer_size);
if (rc != 0) {
CERROR("Can't set buffer tx %d, rx %d buffers: %d\n",
*ksocknal_tunables.ksnd_tx_buffer_size,
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
index 8596581..3916790 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
@@ -495,8 +495,8 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hdr->msg.hello.type = cpu_to_le32 (hello->kshm_ctype);
hdr->msg.hello.incarnation = cpu_to_le64 (hello->kshm_src_incarnation);
- rc = libcfs_sock_write(sock, hdr, sizeof(*hdr),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hdr, sizeof(*hdr),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
@@ -511,9 +511,9 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hello->kshm_ips[i] = __cpu_to_le32 (hello->kshm_ips[i]);
}
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -544,8 +544,8 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
LNET_UNLOCK();
}
- rc = libcfs_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
@@ -556,9 +556,9 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -583,7 +583,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
return -ENOMEM;
}
- rc = libcfs_sock_read(sock, &hdr->src_nid,
+ rc = lnet_sock_read(sock, &hdr->src_nid,
sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
timeout);
if (rc != 0) {
@@ -619,7 +619,7 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
if (hello->kshm_nips == 0)
goto out;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
+ rc = lnet_sock_read(sock, hello->kshm_ips,
hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
@@ -656,7 +656,7 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
else
conn->ksnc_flip = 1;
- rc = libcfs_sock_read(sock, &hello->kshm_src_nid,
+ rc = lnet_sock_read(sock, &hello->kshm_src_nid,
offsetof(ksock_hello_msg_t, kshm_ips) -
offsetof(ksock_hello_msg_t, kshm_src_nid),
timeout);
@@ -687,8 +687,8 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/Makefile b/drivers/staging/lustre/lnet/lnet/Makefile
index 336b8ea..52492fb 100644
--- a/drivers/staging/lustre/lnet/lnet/Makefile
+++ b/drivers/staging/lustre/lnet/lnet/Makefile
@@ -1,5 +1,6 @@
obj-$(CONFIG_LNET) += lnet.o
-lnet-y := api-ni.o config.o lib-me.o lib-msg.o lib-eq.o \
- lib-md.o lib-ptl.o lib-move.o module.o lo.o router.o \
- router_proc.o acceptor.o peer.o
+lnet-y := api-ni.o config.o \
+ lib-me.o lib-msg.o lib-eq.o lib-md.o lib-ptl.o \
+ lib-socket.o lib-move.o module.o lo.o \
+ router.o router_proc.o acceptor.o peer.o
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 72fd1bf..bd97099 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -155,9 +155,8 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
--port) {
/* Iterate through reserved ports. */
- rc = libcfs_sock_connect(&sock, &fatal,
- local_ip, port,
- peer_ip, peer_port);
+ rc = lnet_sock_connect(&sock, &fatal, local_ip, port,
+ peer_ip, peer_port);
if (rc != 0) {
if (fatal)
goto failed;
@@ -184,8 +183,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
lnet_net_unlock(LNET_LOCK_EX);
}
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
goto failed_sock;
@@ -197,7 +195,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
goto failed;
failed_sock:
- libcfs_sock_release(sock);
+ sock_release(sock);
failed:
lnet_connect_console_error(rc, peer_nid, peer_ip, peer_port);
return rc;
@@ -220,7 +218,7 @@ lnet_accept(struct socket *sock, __u32 magic)
LASSERT(sizeof(cr) <= 16); /* not too big for the stack */
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
if (!lnet_accept_magic(magic, LNET_PROTO_ACCEPTOR_MAGIC)) {
@@ -234,7 +232,7 @@ lnet_accept(struct socket *sock, __u32 magic)
memset(&cr, 0, sizeof(cr));
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
accept_timeout);
if (rc != 0)
@@ -257,9 +255,8 @@ lnet_accept(struct socket *sock, __u32 magic)
flip = (magic != LNET_PROTO_ACCEPTOR_MAGIC);
- rc = libcfs_sock_read(sock, &cr.acr_version,
- sizeof(cr.acr_version),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_version, sizeof(cr.acr_version),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request version from %pI4h\n",
rc, &peer_ip);
@@ -280,7 +277,7 @@ lnet_accept(struct socket *sock, __u32 magic)
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
accept_timeout);
if (rc != 0)
@@ -289,7 +286,7 @@ lnet_accept(struct socket *sock, __u32 magic)
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &cr.acr_nid,
+ rc = lnet_sock_read(sock, &cr.acr_nid,
sizeof(cr) -
offsetof(lnet_acceptor_connreq_t, acr_nid),
accept_timeout);
@@ -343,8 +340,8 @@ lnet_acceptor(void *arg)
cfs_block_allsigs();
- rc = libcfs_sock_listen(&lnet_acceptor_state.pta_sock,
- 0, accept_port, accept_backlog);
+ rc = lnet_sock_listen(&lnet_acceptor_state.pta_sock, 0, accept_port,
+ accept_backlog);
if (rc != 0) {
if (rc == -EADDRINUSE)
LCONSOLE_ERROR_MSG(0x122, "Can't start acceptor on port %d: port already in use\n",
@@ -367,7 +364,7 @@ lnet_acceptor(void *arg)
while (!lnet_acceptor_state.pta_shutdown) {
- rc = libcfs_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
+ rc = lnet_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
if (rc != 0) {
if (rc != -EAGAIN) {
CWARN("Accept error %d: pausing...\n", rc);
@@ -377,13 +374,13 @@ lnet_acceptor(void *arg)
continue;
}
- /* maybe we're waken up with libcfs_sock_abort_accept() */
+ /* maybe we're waken up with lnet_sock_abort_accept() */
if (lnet_acceptor_state.pta_shutdown) {
- libcfs_sock_release(newsock);
+ sock_release(newsock);
break;
}
- rc = libcfs_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
if (rc != 0) {
CERROR("Can't determine new connection's address\n");
goto failed;
@@ -395,7 +392,7 @@ lnet_acceptor(void *arg)
goto failed;
}
- rc = libcfs_sock_read(newsock, &magic, sizeof(magic),
+ rc = lnet_sock_read(newsock, &magic, sizeof(magic),
accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
@@ -410,10 +407,10 @@ lnet_acceptor(void *arg)
continue;
failed:
- libcfs_sock_release(newsock);
+ sock_release(newsock);
}
- libcfs_sock_release(lnet_acceptor_state.pta_sock);
+ sock_release(lnet_acceptor_state.pta_sock);
lnet_acceptor_state.pta_sock = NULL;
CDEBUG(D_NET, "Acceptor stopping\n");
@@ -493,7 +490,7 @@ lnet_acceptor_stop(void)
return;
lnet_acceptor_state.pta_shutdown = 1;
- libcfs_sock_abort_accept(lnet_acceptor_state.pta_sock);
+ wake_up_all(sk_sleep(lnet_acceptor_state.pta_sock->sk));
/* block until acceptor signals exit */
wait_for_completion(&lnet_acceptor_state.pta_signal);
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index 2dc4c4a..d8c8f2e 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -1118,7 +1118,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
__u32 *ipaddrs2;
int nip;
char **ifnames;
- int nif = libcfs_ipif_enumerate(&ifnames);
+ int nif = lnet_ipif_enumerate(&ifnames);
int i;
int rc;
@@ -1128,7 +1128,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
LIBCFS_ALLOC(ipaddrs, nif * sizeof(*ipaddrs));
if (ipaddrs == NULL) {
CERROR("Can't allocate ipaddrs[%d]\n", nif);
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
return -ENOMEM;
}
@@ -1136,8 +1136,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
if (!strcmp(ifnames[i], "lo"))
continue;
- rc = libcfs_ipif_query(ifnames[i], &up,
- &ipaddrs[nip], &netmask);
+ rc = lnet_ipif_query(ifnames[i], &up, &ipaddrs[nip], &netmask);
if (rc != 0) {
CWARN("Can't query interface %s: %d\n",
ifnames[i], rc);
@@ -1153,7 +1152,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
nip++;
}
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
if (nip == nif) {
*ipaddrsp = ipaddrs;
@@ -1237,8 +1236,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return -EPERM;
}
- rc = libcfs_ipif_query(ni->ni_interfaces[0],
- &up, &ip, &netmask);
+ rc = lnet_ipif_query(ni->ni_interfaces[0], &up, &ip, &netmask);
if (rc != 0) {
CERROR("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), ni->ni_interfaces[0], rc);
@@ -1255,7 +1253,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return 0;
}
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Net %s can't enumerate interfaces: %d\n",
libcfs_net2str(net), n);
@@ -1266,8 +1264,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &netmask);
-
+ rc = lnet_ipif_query(names[i], &up, &ip, &netmask);
if (rc != 0) {
CWARN("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), names[i], rc);
@@ -1280,13 +1277,13 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
continue;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
ni->ni_nid = LNET_MKNID(net, ip);
return 0;
}
CERROR("Net %s can't find any interfaces\n", libcfs_net2str(net));
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
return -ENOENT;
}
EXPORT_SYMBOL(lnet_set_ip_niaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
new file mode 100644
index 0000000..da7b871
--- /dev/null
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -0,0 +1,616 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, 2014, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+#define DEBUG_SUBSYSTEM S_LNET
+
+#include <linux/if.h>
+#include <linux/in.h>
+#include <linux/net.h>
+#include <linux/file.h>
+#include <linux/pagemap.h>
+/* For sys_open & sys_close */
+#include <linux/syscalls.h>
+#include <net/sock.h>
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/lib-lnet.h"
+
+static int
+kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
+{
+ mm_segment_t oldfs = get_fs();
+ int err;
+
+ set_fs(KERNEL_DS);
+ err = filp->f_op->unlocked_ioctl(filp, cmd, arg);
+ set_fs(oldfs);
+
+ return err;
+}
+
+static int
+lnet_sock_ioctl(int cmd, unsigned long arg)
+{
+ struct file *sock_filp;
+ struct socket *sock;
+ int fd = -1;
+ int rc;
+
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
+ if (rc != 0) {
+ CERROR("Can't create socket: %d\n", rc);
+ return rc;
+ }
+
+ sock_filp = sock_alloc_file(sock, 0, NULL);
+ if (!sock_filp) {
+ rc = -ENOMEM;
+ sock_release(sock);
+ goto out;
+ }
+
+ rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
+
+ fput(sock_filp);
+out:
+ if (fd >= 0)
+ sys_close(fd);
+ return rc;
+}
+
+int
+lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask)
+{
+ struct ifreq ifr;
+ int nob;
+ int rc;
+ __u32 val;
+
+ nob = strnlen(name, IFNAMSIZ);
+ if (nob == IFNAMSIZ) {
+ CERROR("Interface name %s too long\n", name);
+ return -EINVAL;
+ }
+
+ CLASSERT(sizeof(ifr.ifr_name) >= IFNAMSIZ);
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
+ if (rc != 0) {
+ CERROR("Can't get flags for interface %s\n", name);
+ return rc;
+ }
+
+ if ((ifr.ifr_flags & IFF_UP) == 0) {
+ CDEBUG(D_NET, "Interface %s down\n", name);
+ *up = 0;
+ *ip = *mask = 0;
+ return 0;
+ }
+ *up = 1;
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ ifr.ifr_addr.sa_family = AF_INET;
+ rc = lnet_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
+
+ if (rc != 0) {
+ CERROR("Can't get IP address for interface %s\n", name);
+ return rc;
+ }
+
+ val = ((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr;
+ *ip = ntohl(val);
+
+ if (strlen(name) > sizeof(ifr.ifr_name)-1)
+ return -E2BIG;
+ strncpy(ifr.ifr_name, name, sizeof(ifr.ifr_name));
+
+ ifr.ifr_addr.sa_family = AF_INET;
+ rc = lnet_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
+ if (rc != 0) {
+ CERROR("Can't get netmask for interface %s\n", name);
+ return rc;
+ }
+
+ val = ((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr.s_addr;
+ *mask = ntohl(val);
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_ipif_query);
+
+void
+lnet_ipif_free_enumeration(char **names, int n)
+{
+ int i;
+
+ LASSERT(n > 0);
+
+ for (i = 0; i < n && names[i] != NULL; i++)
+ LIBCFS_FREE(names[i], IFNAMSIZ);
+
+ LIBCFS_FREE(names, n * sizeof(*names));
+}
+EXPORT_SYMBOL(lnet_ipif_free_enumeration);
+
+int
+lnet_ipif_enumerate(char ***namesp)
+{
+ /* Allocate and fill in 'names', returning # interfaces/error */
+ char **names;
+ int toobig;
+ int nalloc;
+ int nfound;
+ struct ifreq *ifr;
+ struct ifconf ifc;
+ int rc;
+ int nob;
+ int i;
+
+ nalloc = 16; /* first guess at max interfaces */
+ toobig = 0;
+ for (;;) {
+ if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
+ toobig = 1;
+ nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
+ CWARN("Too many interfaces: only enumerating "
+ "first %d\n", nalloc);
+ }
+
+ LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
+ if (ifr == NULL) {
+ CERROR("ENOMEM enumerating up to %d interfaces\n",
+ nalloc);
+ rc = -ENOMEM;
+ goto out0;
+ }
+
+ ifc.ifc_buf = (char *)ifr;
+ ifc.ifc_len = nalloc * sizeof(*ifr);
+
+ rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
+ if (rc < 0) {
+ CERROR("Error %d enumerating interfaces\n", rc);
+ goto out1;
+ }
+
+ LASSERT(rc == 0);
+
+ nfound = ifc.ifc_len/sizeof(*ifr);
+ LASSERT(nfound <= nalloc);
+
+ if (nfound < nalloc || toobig)
+ break;
+
+ LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
+ nalloc *= 2;
+ }
+
+ if (nfound == 0)
+ goto out1;
+
+ LIBCFS_ALLOC(names, nfound * sizeof(*names));
+ if (names == NULL) {
+ rc = -ENOMEM;
+ goto out1;
+ }
+
+ for (i = 0; i < nfound; i++) {
+ nob = strnlen(ifr[i].ifr_name, IFNAMSIZ);
+ if (nob == IFNAMSIZ) {
+ /* no space for terminating NULL */
+ CERROR("interface name %.*s too long (%d max)\n",
+ nob, ifr[i].ifr_name, IFNAMSIZ);
+ rc = -ENAMETOOLONG;
+ goto out2;
+ }
+
+ LIBCFS_ALLOC(names[i], IFNAMSIZ);
+ if (names[i] == NULL) {
+ rc = -ENOMEM;
+ goto out2;
+ }
+
+ memcpy(names[i], ifr[i].ifr_name, nob);
+ names[i][nob] = 0;
+ }
+
+ *namesp = names;
+ rc = nfound;
+
+ out2:
+ if (rc < 0)
+ lnet_ipif_free_enumeration(names, nfound);
+ out1:
+ LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
+ out0:
+ return rc;
+}
+EXPORT_SYMBOL(lnet_ipif_enumerate);
+
+int
+lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout)
+{
+ int rc;
+ long jiffies_left = timeout * msecs_to_jiffies(MSEC_PER_SEC);
+ unsigned long then;
+ struct timeval tv;
+
+ LASSERT(nob > 0);
+ /* Caller may pass a zero timeout if she thinks the socket buffer is
+ * empty enough to take the whole message immediately */
+
+ for (;;) {
+ struct kvec iov = {
+ .iov_base = buffer,
+ .iov_len = nob
+ };
+ struct msghdr msg = {
+ .msg_flags = (timeout == 0) ? MSG_DONTWAIT : 0
+ };
+
+ if (timeout != 0) {
+ /* Set send timeout to remaining time */
+ tv = (struct timeval) {
+ .tv_sec = jiffies_left /
+ msecs_to_jiffies(MSEC_PER_SEC),
+ .tv_usec = ((jiffies_left %
+ msecs_to_jiffies(MSEC_PER_SEC)) *
+ USEC_PER_SEC) /
+ msecs_to_jiffies(MSEC_PER_SEC)
+ };
+
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
+ (char *)&tv, sizeof(tv));
+ if (rc != 0) {
+ CERROR("Can't set socket send timeout "
+ "%ld.%06d: %d\n",
+ (long)tv.tv_sec, (int)tv.tv_usec, rc);
+ return rc;
+ }
+ }
+
+ then = jiffies;
+ rc = kernel_sendmsg(sock, &msg, &iov, 1, nob);
+ jiffies_left -= jiffies - then;
+
+ if (rc == nob)
+ return 0;
+
+ if (rc < 0)
+ return rc;
+
+ if (rc == 0) {
+ CERROR("Unexpected zero rc\n");
+ return -ECONNABORTED;
+ }
+
+ if (jiffies_left <= 0)
+ return -EAGAIN;
+
+ buffer = ((char *)buffer) + rc;
+ nob -= rc;
+ }
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_write);
+
+int
+lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout)
+{
+ int rc;
+ long jiffies_left = timeout * msecs_to_jiffies(MSEC_PER_SEC);
+ unsigned long then;
+ struct timeval tv;
+
+ LASSERT(nob > 0);
+ LASSERT(jiffies_left > 0);
+
+ for (;;) {
+ struct kvec iov = {
+ .iov_base = buffer,
+ .iov_len = nob
+ };
+ struct msghdr msg = {
+ .msg_flags = 0
+ };
+
+ /* Set receive timeout to remaining time */
+ tv = (struct timeval) {
+ .tv_sec = jiffies_left / msecs_to_jiffies(MSEC_PER_SEC),
+ .tv_usec = ((jiffies_left %
+ msecs_to_jiffies(MSEC_PER_SEC)) *
+ USEC_PER_SEC) /
+ msecs_to_jiffies(MSEC_PER_SEC)
+ };
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
+ (char *)&tv, sizeof(tv));
+ if (rc != 0) {
+ CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
+ (long)tv.tv_sec, (int)tv.tv_usec, rc);
+ return rc;
+ }
+
+ then = jiffies;
+ rc = kernel_recvmsg(sock, &msg, &iov, 1, nob, 0);
+ jiffies_left -= jiffies - then;
+
+ if (rc < 0)
+ return rc;
+
+ if (rc == 0)
+ return -ECONNRESET;
+
+ buffer = ((char *)buffer) + rc;
+ nob -= rc;
+
+ if (nob == 0)
+ return 0;
+
+ if (jiffies_left <= 0)
+ return -ETIMEDOUT;
+ }
+}
+EXPORT_SYMBOL(lnet_sock_read);
+
+static int
+lnet_sock_create(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port)
+{
+ struct sockaddr_in locaddr;
+ struct socket *sock;
+ int rc;
+ int option;
+
+ /* All errors are fatal except bind failure if the port is in use */
+ *fatal = 1;
+
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
+ *sockp = sock;
+ if (rc != 0) {
+ CERROR("Can't create socket: %d\n", rc);
+ return rc;
+ }
+
+ option = 1;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
+ goto failed;
+ }
+
+ if (local_ip != 0 || local_port != 0) {
+ memset(&locaddr, 0, sizeof(locaddr));
+ locaddr.sin_family = AF_INET;
+ locaddr.sin_port = htons(local_port);
+ locaddr.sin_addr.s_addr = (local_ip == 0) ?
+ INADDR_ANY : htonl(local_ip);
+
+ rc = kernel_bind(sock, (struct sockaddr *)&locaddr,
+ sizeof(locaddr));
+ if (rc == -EADDRINUSE) {
+ CDEBUG(D_NET, "Port %d already in use\n", local_port);
+ *fatal = 0;
+ goto failed;
+ }
+ if (rc != 0) {
+ CERROR("Error trying to bind to port %d: %d\n",
+ local_port, rc);
+ goto failed;
+ }
+ }
+ return 0;
+
+failed:
+ sock_release(sock);
+ return rc;
+}
+
+int
+lnet_sock_setbuf(struct socket *sock, int txbufsize, int rxbufsize)
+{
+ int option;
+ int rc;
+
+ if (txbufsize != 0) {
+ option = txbufsize;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set send buffer %d: %d\n",
+ option, rc);
+ return rc;
+ }
+ }
+
+ if (rxbufsize != 0) {
+ option = rxbufsize;
+ rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
+ (char *)&option, sizeof(option));
+ if (rc != 0) {
+ CERROR("Can't set receive buffer %d: %d\n",
+ option, rc);
+ return rc;
+ }
+ }
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_setbuf);
+
+int
+lnet_sock_getaddr(struct socket *sock, bool remote, __u32 *ip, int *port)
+{
+ struct sockaddr_in sin;
+ int len = sizeof(sin);
+ int rc;
+
+ if (remote)
+ rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
+ else
+ rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
+ if (rc != 0) {
+ CERROR("Error %d getting sock %s IP/port\n",
+ rc, remote ? "peer" : "local");
+ return rc;
+ }
+
+ if (ip != NULL)
+ *ip = ntohl(sin.sin_addr.s_addr);
+
+ if (port != NULL)
+ *port = ntohs(sin.sin_port);
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_getaddr);
+
+int
+lnet_sock_getbuf(struct socket *sock, int *txbufsize, int *rxbufsize)
+{
+ if (txbufsize != NULL)
+ *txbufsize = sock->sk->sk_sndbuf;
+
+ if (rxbufsize != NULL)
+ *rxbufsize = sock->sk->sk_rcvbuf;
+
+ return 0;
+}
+EXPORT_SYMBOL(lnet_sock_getbuf);
+
+int
+lnet_sock_listen(struct socket **sockp,
+ __u32 local_ip, int local_port, int backlog)
+{
+ int fatal;
+ int rc;
+
+ rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
+ if (rc != 0) {
+ if (!fatal)
+ CERROR("Can't create socket: port %d already in use\n",
+ local_port);
+ return rc;
+ }
+
+ rc = kernel_listen(*sockp, backlog);
+ if (rc == 0)
+ return 0;
+
+ CERROR("Can't set listen backlog %d: %d\n", backlog, rc);
+ sock_release(*sockp);
+ return rc;
+}
+
+int
+lnet_sock_accept(struct socket **newsockp, struct socket *sock)
+{
+ wait_queue_t wait;
+ struct socket *newsock;
+ int rc;
+
+ /* XXX this should add a ref to sock->ops->owner, if
+ * TCP could be a module */
+ rc = sock_create_lite(PF_PACKET, sock->type, IPPROTO_TCP, &newsock);
+ if (rc) {
+ CERROR("Can't allocate socket\n");
+ return rc;
+ }
+
+ newsock->ops = sock->ops;
+
+ rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
+ if (rc == -EAGAIN) {
+ /* Nothing ready, so wait for activity */
+ init_waitqueue_entry(&wait, current);
+ add_wait_queue(sk_sleep(sock->sk), &wait);
+ set_current_state(TASK_INTERRUPTIBLE);
+ schedule();
+ remove_wait_queue(sk_sleep(sock->sk), &wait);
+ rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
+ }
+
+ if (rc != 0)
+ goto failed;
+
+ *newsockp = newsock;
+ return 0;
+
+failed:
+ sock_release(newsock);
+ return rc;
+}
+
+int
+lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port)
+{
+ struct sockaddr_in srvaddr;
+ int rc;
+
+ rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
+ if (rc != 0)
+ return rc;
+
+ memset(&srvaddr, 0, sizeof(srvaddr));
+ srvaddr.sin_family = AF_INET;
+ srvaddr.sin_port = htons(peer_port);
+ srvaddr.sin_addr.s_addr = htonl(peer_ip);
+
+ rc = kernel_connect(*sockp, (struct sockaddr *)&srvaddr,
+ sizeof(srvaddr), 0);
+ if (rc == 0)
+ return 0;
+
+ /* EADDRNOTAVAIL probably means we're already connected to the same
+ * peer/port on the same local port on a differently typed
+ * connection. Let our caller retry with a different local
+ * port... */
+ *fatal = !(rc == -EADDRNOTAVAIL);
+
+ CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
+ "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
+ &local_ip, local_port, &peer_ip, peer_port);
+
+ sock_release(*sockp);
+ return rc;
+}
diff --git a/drivers/staging/lustre/lustre/libcfs/Makefile b/drivers/staging/lustre/lustre/libcfs/Makefile
index fabdd3e..ec98f44 100644
--- a/drivers/staging/lustre/lustre/libcfs/Makefile
+++ b/drivers/staging/lustre/lustre/libcfs/Makefile
@@ -2,7 +2,6 @@ obj-$(CONFIG_LUSTRE_FS) += libcfs.o
libcfs-linux-objs := linux-tracefile.o linux-debug.o
libcfs-linux-objs += linux-prim.o linux-cpu.o
-libcfs-linux-objs += linux-tcpip.o
libcfs-linux-objs += linux-curproc.o
libcfs-linux-objs += linux-module.o
libcfs-linux-objs += linux-crypto.o
diff --git a/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c b/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
deleted file mode 100644
index f2462e7..0000000
--- a/drivers/staging/lustre/lustre/libcfs/linux/linux-tcpip.c
+++ /dev/null
@@ -1,623 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2012, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-#define DEBUG_SUBSYSTEM S_LNET
-
-#include "../../../include/linux/libcfs/libcfs.h"
-
-#include <linux/if.h>
-#include <linux/in.h>
-#include <linux/file.h>
-/* For sys_open & sys_close */
-#include <linux/syscalls.h>
-
-static int
-libcfs_sock_ioctl(int cmd, unsigned long arg)
-{
- mm_segment_t oldmm = get_fs();
- struct socket *sock;
- int rc;
- struct file *sock_filp;
-
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
- if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
- return rc;
- }
-
- sock_filp = sock_alloc_file(sock, 0, NULL);
- if (IS_ERR(sock_filp)) {
- sock_release(sock);
- rc = PTR_ERR(sock_filp);
- goto out;
- }
-
- set_fs(KERNEL_DS);
- if (sock_filp->f_op->unlocked_ioctl)
- rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
- set_fs(oldmm);
-
- fput(sock_filp);
-out:
- return rc;
-}
-
-int
-libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
-{
- struct ifreq ifr;
- int nob;
- int rc;
- __u32 val;
-
- nob = strnlen(name, IFNAMSIZ);
- if (nob == IFNAMSIZ) {
- CERROR("Interface name %s too long\n", name);
- return -EINVAL;
- }
-
- CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
-
- strcpy(ifr.ifr_name, name);
- rc = libcfs_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get flags for interface %s\n", name);
- return rc;
- }
-
- if ((ifr.ifr_flags & IFF_UP) == 0) {
- CDEBUG(D_NET, "Interface %s down\n", name);
- *up = 0;
- *ip = *mask = 0;
- return 0;
- }
-
- *up = 1;
-
- strcpy(ifr.ifr_name, name);
- ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get IP address for interface %s\n", name);
- return rc;
- }
-
- val = ((struct sockaddr_in *)&ifr.ifr_addr)->sin_addr.s_addr;
- *ip = ntohl(val);
-
- strcpy(ifr.ifr_name, name);
- ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
-
- if (rc != 0) {
- CERROR("Can't get netmask for interface %s\n", name);
- return rc;
- }
-
- val = ((struct sockaddr_in *)&ifr.ifr_netmask)->sin_addr.s_addr;
- *mask = ntohl(val);
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_ipif_query);
-
-int
-libcfs_ipif_enumerate (char ***namesp)
-{
- /* Allocate and fill in 'names', returning # interfaces/error */
- char **names;
- int toobig;
- int nalloc;
- int nfound;
- struct ifreq *ifr;
- struct ifconf ifc;
- int rc;
- int nob;
- int i;
-
-
- nalloc = 16; /* first guess at max interfaces */
- toobig = 0;
- for (;;) {
- if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
- toobig = 1;
- nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
- CWARN("Too many interfaces: only enumerating first %d\n",
- nalloc);
- }
-
- LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
- if (ifr == NULL) {
- CERROR ("ENOMEM enumerating up to %d interfaces\n", nalloc);
- rc = -ENOMEM;
- goto out0;
- }
-
- ifc.ifc_buf = (char *)ifr;
- ifc.ifc_len = nalloc * sizeof(*ifr);
-
- rc = libcfs_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
-
- if (rc < 0) {
- CERROR ("Error %d enumerating interfaces\n", rc);
- goto out1;
- }
-
- LASSERT (rc == 0);
-
- nfound = ifc.ifc_len/sizeof(*ifr);
- LASSERT (nfound <= nalloc);
-
- if (nfound < nalloc || toobig)
- break;
-
- LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- nalloc *= 2;
- }
-
- if (nfound == 0)
- goto out1;
-
- LIBCFS_ALLOC(names, nfound * sizeof(*names));
- if (names == NULL) {
- rc = -ENOMEM;
- goto out1;
- }
-
- for (i = 0; i < nfound; i++) {
-
- nob = strnlen (ifr[i].ifr_name, IFNAMSIZ);
- if (nob == IFNAMSIZ) {
- /* no space for terminating NULL */
- CERROR("interface name %.*s too long (%d max)\n",
- nob, ifr[i].ifr_name, IFNAMSIZ);
- rc = -ENAMETOOLONG;
- goto out2;
- }
-
- LIBCFS_ALLOC(names[i], IFNAMSIZ);
- if (names[i] == NULL) {
- rc = -ENOMEM;
- goto out2;
- }
-
- memcpy(names[i], ifr[i].ifr_name, nob);
- names[i][nob] = 0;
- }
-
- *namesp = names;
- rc = nfound;
-
- out2:
- if (rc < 0)
- libcfs_ipif_free_enumeration(names, nfound);
- out1:
- LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- out0:
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_ipif_enumerate);
-
-void
-libcfs_ipif_free_enumeration (char **names, int n)
-{
- int i;
-
- LASSERT (n > 0);
-
- for (i = 0; i < n && names[i] != NULL; i++)
- LIBCFS_FREE(names[i], IFNAMSIZ);
-
- LIBCFS_FREE(names, n * sizeof(*names));
-}
-
-EXPORT_SYMBOL(libcfs_ipif_free_enumeration);
-
-int
-libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
-{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
- struct timeval tv;
-
- LASSERT (nob > 0);
- /* Caller may pass a zero timeout if she thinks the socket buffer is
- * empty enough to take the whole message immediately */
-
- for (;;) {
- struct kvec iov = {
- .iov_base = buffer,
- .iov_len = nob
- };
- struct msghdr msg = {
- .msg_flags = (timeout == 0) ? MSG_DONTWAIT : 0
- };
-
- if (timeout != 0) {
- /* Set send timeout to remaining time */
- tv = (struct timeval) {
- .tv_sec = ticks / HZ,
- .tv_usec = ((ticks % HZ) * 1000000) / HZ
- };
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
- (char *)&tv, sizeof(tv));
- if (rc != 0) {
- CERROR("Can't set socket send timeout %ld.%06d: %d\n",
- (long)tv.tv_sec, (int)tv.tv_usec, rc);
- return rc;
- }
- }
-
- then = jiffies;
- rc = kernel_sendmsg(sock, &msg, &iov, 1, nob);
- ticks -= jiffies - then;
-
- if (rc == nob)
- return 0;
-
- if (rc < 0)
- return rc;
-
- if (rc == 0) {
- CERROR ("Unexpected zero rc\n");
- return -ECONNABORTED;
- }
-
- if (ticks <= 0)
- return -EAGAIN;
-
- buffer = ((char *)buffer) + rc;
- nob -= rc;
- }
-
- return 0;
-}
-EXPORT_SYMBOL(libcfs_sock_write);
-
-int
-libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
-{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
- struct timeval tv;
-
- LASSERT (nob > 0);
- LASSERT (ticks > 0);
-
- for (;;) {
- struct kvec iov = {
- .iov_base = buffer,
- .iov_len = nob
- };
- struct msghdr msg = {
- .msg_flags = 0
- };
-
- /* Set receive timeout to remaining time */
- tv = (struct timeval) {
- .tv_sec = ticks / HZ,
- .tv_usec = ((ticks % HZ) * 1000000) / HZ
- };
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
- (char *)&tv, sizeof(tv));
- if (rc != 0) {
- CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
- (long)tv.tv_sec, (int)tv.tv_usec, rc);
- return rc;
- }
-
- then = jiffies;
- rc = kernel_recvmsg(sock, &msg, &iov, 1, nob, 0);
- ticks -= jiffies - then;
-
- if (rc < 0)
- return rc;
-
- if (rc == 0)
- return -ECONNRESET;
-
- buffer = ((char *)buffer) + rc;
- nob -= rc;
-
- if (nob == 0)
- return 0;
-
- if (ticks <= 0)
- return -ETIMEDOUT;
- }
-}
-
-EXPORT_SYMBOL(libcfs_sock_read);
-
-static int
-libcfs_sock_create (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port)
-{
- struct sockaddr_in locaddr;
- struct socket *sock;
- int rc;
- int option;
-
- /* All errors are fatal except bind failure if the port is in use */
- *fatal = 1;
-
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
- *sockp = sock;
- if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
- return rc;
- }
-
- option = 1;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
- goto failed;
- }
-
- if (local_ip != 0 || local_port != 0) {
- memset(&locaddr, 0, sizeof(locaddr));
- locaddr.sin_family = AF_INET;
- locaddr.sin_port = htons(local_port);
- locaddr.sin_addr.s_addr = (local_ip == 0) ?
- INADDR_ANY : htonl(local_ip);
-
- rc = sock->ops->bind(sock, (struct sockaddr *)&locaddr,
- sizeof(locaddr));
- if (rc == -EADDRINUSE) {
- CDEBUG(D_NET, "Port %d already in use\n", local_port);
- *fatal = 0;
- goto failed;
- }
- if (rc != 0) {
- CERROR("Error trying to bind to port %d: %d\n",
- local_port, rc);
- goto failed;
- }
- }
-
- return 0;
-
- failed:
- sock_release(sock);
- return rc;
-}
-
-int
-libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
-{
- int option;
- int rc;
-
- if (txbufsize != 0) {
- option = txbufsize;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR ("Can't set send buffer %d: %d\n",
- option, rc);
- return rc;
- }
- }
-
- if (rxbufsize != 0) {
- option = rxbufsize;
- rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
- (char *)&option, sizeof (option));
- if (rc != 0) {
- CERROR ("Can't set receive buffer %d: %d\n",
- option, rc);
- return rc;
- }
- }
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_setbuf);
-
-int
-libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
-{
- struct sockaddr_in sin;
- int len = sizeof (sin);
- int rc;
-
- rc = sock->ops->getname (sock, (struct sockaddr *)&sin, &len,
- remote ? 2 : 0);
- if (rc != 0) {
- CERROR ("Error %d getting sock %s IP/port\n",
- rc, remote ? "peer" : "local");
- return rc;
- }
-
- if (ip != NULL)
- *ip = ntohl (sin.sin_addr.s_addr);
-
- if (port != NULL)
- *port = ntohs (sin.sin_port);
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_getaddr);
-
-int
-libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
-{
-
- if (txbufsize != NULL) {
- *txbufsize = sock->sk->sk_sndbuf;
- }
-
- if (rxbufsize != NULL) {
- *rxbufsize = sock->sk->sk_rcvbuf;
- }
-
- return 0;
-}
-
-EXPORT_SYMBOL(libcfs_sock_getbuf);
-
-int
-libcfs_sock_listen (struct socket **sockp,
- __u32 local_ip, int local_port, int backlog)
-{
- int fatal;
- int rc;
-
- rc = libcfs_sock_create(sockp, &fatal, local_ip, local_port);
- if (rc != 0) {
- if (!fatal)
- CERROR("Can't create socket: port %d already in use\n",
- local_port);
- return rc;
- }
-
- rc = (*sockp)->ops->listen(*sockp, backlog);
- if (rc == 0)
- return 0;
-
- CERROR("Can't set listen backlog %d: %d\n", backlog, rc);
- sock_release(*sockp);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_listen);
-
-int
-libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
-{
- wait_queue_t wait;
- struct socket *newsock;
- int rc;
-
- init_waitqueue_entry(&wait, current);
-
- /* XXX this should add a ref to sock->ops->owner, if
- * TCP could be a module */
- rc = sock_create_lite(PF_PACKET, sock->type, IPPROTO_TCP, &newsock);
- if (rc) {
- CERROR("Can't allocate socket\n");
- return rc;
- }
-
- newsock->ops = sock->ops;
-
- rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
- if (rc == -EAGAIN) {
- /* Nothing ready, so wait for activity */
- set_current_state(TASK_INTERRUPTIBLE);
- add_wait_queue(sk_sleep(sock->sk), &wait);
- schedule();
- remove_wait_queue(sk_sleep(sock->sk), &wait);
- set_current_state(TASK_RUNNING);
- rc = sock->ops->accept(sock, newsock, O_NONBLOCK);
- }
-
- if (rc != 0)
- goto failed;
-
- *newsockp = newsock;
- return 0;
-
- failed:
- sock_release(newsock);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_accept);
-
-void
-libcfs_sock_abort_accept (struct socket *sock)
-{
- wake_up_all(sk_sleep(sock->sk));
-}
-
-EXPORT_SYMBOL(libcfs_sock_abort_accept);
-
-int
-libcfs_sock_connect (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port)
-{
- struct sockaddr_in srvaddr;
- int rc;
-
- rc = libcfs_sock_create(sockp, fatal, local_ip, local_port);
- if (rc != 0)
- return rc;
-
- memset (&srvaddr, 0, sizeof (srvaddr));
- srvaddr.sin_family = AF_INET;
- srvaddr.sin_port = htons(peer_port);
- srvaddr.sin_addr.s_addr = htonl(peer_ip);
-
- rc = (*sockp)->ops->connect(*sockp,
- (struct sockaddr *)&srvaddr, sizeof(srvaddr),
- 0);
- if (rc == 0)
- return 0;
-
- /* EADDRNOTAVAIL probably means we're already connected to the same
- * peer/port on the same local port on a differently typed
- * connection. Let our caller retry with a different local
- * port... */
- *fatal = !(rc == -EADDRNOTAVAIL);
-
- CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
- "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
- &local_ip, local_port, &peer_ip, peer_port);
-
- sock_release(*sockp);
- return rc;
-}
-
-EXPORT_SYMBOL(libcfs_sock_connect);
-
-void
-libcfs_sock_release (struct socket *sock)
-{
- sock_release(sock);
-}
-
-EXPORT_SYMBOL(libcfs_sock_release);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
2015-05-21 22:46 ` James Simmons
@ 2015-05-22 11:15 ` Dan Carpenter
-1 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 11:15 UTC (permalink / raw)
To: lustre-devel
This patch does a lot of stuff all at once and it is hard to review. It
could easily be broken into patches which are easy to review.
> @@ -1378,15 +1378,15 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
> ksocknal_txlist_done(ni, &zombies, 1);
> ksocknal_peer_decref(peer);
>
> - failed_1:
> +failed_1:
Do unrelated white space changes in a different patch. It just makes
reviewing complicated to mix easy to review white space changes in with
everything else.
> if (hello != NULL)
> LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
> kshm_ips[LNET_MAX_INTERFACES]));
>
> LIBCFS_FREE(conn, sizeof(*conn));
>
> - failed_0:
> - libcfs_sock_release(sock);
> +failed_0:
> + sock_release(sock);
Do a rename patch by itself. You can rename a bunch of functions at the
same time, that's fine. Personally, you can even move the functions
between files, that's also fine with me and doesn't complicate my
review.
Ok in this next section we move functions around and rename them but
also introduce some bad changes in the new function.
> +static int
> +lnet_sock_ioctl(int cmd, unsigned long arg)
> +{
> + struct file *sock_filp;
> + struct socket *sock;
> + int fd = -1;
> + int rc;
> +
> + rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
> + if (rc != 0) {
> + CERROR("Can't create socket: %d\n", rc);
> + return rc;
> + }
> +
> + sock_filp = sock_alloc_file(sock, 0, NULL);
> + if (!sock_filp) {
sock_alloc_file() never returns NULL, only valid pointers on success or
ERR_PTRs on failue.
> + rc = -ENOMEM;
> + sock_release(sock);
> + goto out;
> + }
> +
> + rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
This is an unrelated cleanup. Do it in a different patch.
> +
> + fput(sock_filp);
> +out:
> + if (fd >= 0)
> + sys_close(fd);
This is a new change as well. "fd" is always -1 so this is dead code.
> + return rc;
> +}
> +
Here is the old function:
> -static int
> -libcfs_sock_ioctl(int cmd, unsigned long arg)
> -{
> - mm_segment_t oldmm = get_fs();
> - struct socket *sock;
> - int rc;
> - struct file *sock_filp;
> -
> - rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
> - if (rc != 0) {
> - CERROR ("Can't create socket: %d\n", rc);
> - return rc;
> - }
> -
> - sock_filp = sock_alloc_file(sock, 0, NULL);
> - if (IS_ERR(sock_filp)) {
This check was correct in the original.
> - sock_release(sock);
> - rc = PTR_ERR(sock_filp);
> - goto out;
> - }
> -
> - set_fs(KERNEL_DS);
> - if (sock_filp->f_op->unlocked_ioctl)
> - rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
> - set_fs(oldmm);
> -
> - fput(sock_filp);
> -out:
> - return rc;
> -}
This patch has some bug fixes as well. Those should be sent as one
patch per bugfix with a proper changelog.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* Re: [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 11:15 ` Dan Carpenter
0 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 11:15 UTC (permalink / raw)
To: James Simmons
Cc: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger,
HPDD-discuss, James Simmons, James Simmons,
Linux Kernel Mailing List, lustre-devel
This patch does a lot of stuff all at once and it is hard to review. It
could easily be broken into patches which are easy to review.
> @@ -1378,15 +1378,15 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
> ksocknal_txlist_done(ni, &zombies, 1);
> ksocknal_peer_decref(peer);
>
> - failed_1:
> +failed_1:
Do unrelated white space changes in a different patch. It just makes
reviewing complicated to mix easy to review white space changes in with
everything else.
> if (hello != NULL)
> LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
> kshm_ips[LNET_MAX_INTERFACES]));
>
> LIBCFS_FREE(conn, sizeof(*conn));
>
> - failed_0:
> - libcfs_sock_release(sock);
> +failed_0:
> + sock_release(sock);
Do a rename patch by itself. You can rename a bunch of functions at the
same time, that's fine. Personally, you can even move the functions
between files, that's also fine with me and doesn't complicate my
review.
Ok in this next section we move functions around and rename them but
also introduce some bad changes in the new function.
> +static int
> +lnet_sock_ioctl(int cmd, unsigned long arg)
> +{
> + struct file *sock_filp;
> + struct socket *sock;
> + int fd = -1;
> + int rc;
> +
> + rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
> + if (rc != 0) {
> + CERROR("Can't create socket: %d\n", rc);
> + return rc;
> + }
> +
> + sock_filp = sock_alloc_file(sock, 0, NULL);
> + if (!sock_filp) {
sock_alloc_file() never returns NULL, only valid pointers on success or
ERR_PTRs on failue.
> + rc = -ENOMEM;
> + sock_release(sock);
> + goto out;
> + }
> +
> + rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
This is an unrelated cleanup. Do it in a different patch.
> +
> + fput(sock_filp);
> +out:
> + if (fd >= 0)
> + sys_close(fd);
This is a new change as well. "fd" is always -1 so this is dead code.
> + return rc;
> +}
> +
Here is the old function:
> -static int
> -libcfs_sock_ioctl(int cmd, unsigned long arg)
> -{
> - mm_segment_t oldmm = get_fs();
> - struct socket *sock;
> - int rc;
> - struct file *sock_filp;
> -
> - rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
> - if (rc != 0) {
> - CERROR ("Can't create socket: %d\n", rc);
> - return rc;
> - }
> -
> - sock_filp = sock_alloc_file(sock, 0, NULL);
> - if (IS_ERR(sock_filp)) {
This check was correct in the original.
> - sock_release(sock);
> - rc = PTR_ERR(sock_filp);
> - goto out;
> - }
> -
> - set_fs(KERNEL_DS);
> - if (sock_filp->f_op->unlocked_ioctl)
> - rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
> - set_fs(oldmm);
> -
> - fput(sock_filp);
> -out:
> - return rc;
> -}
This patch has some bug fixes as well. Those should be sent as one
patch per bugfix with a proper changelog.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
2015-05-22 11:15 ` Dan Carpenter
@ 2015-05-22 15:08 ` Simmons, James A.
-1 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-22 15:08 UTC (permalink / raw)
To: lustre-devel
>This patch does a lot of stuff all at once and it is hard to review. It
>could easily be broken into patches which are easy to review.
I have more very large patches. With breaking them up that means you are
going to see hundreds of patches coming from me.
>Ok in this next section we move functions around and rename them but
>also introduce some bad changes in the new function.
Thanks for pointing out these bugs. We are still carrying these bugs in the
OpenSFS branch. Once you approve these changes I will sync up lib-socket.c
in the OpenSFS branch.
P.S
Does the 3rd patch look okay to you?
^ permalink raw reply [flat|nested] 44+ messages in thread* RE: [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 15:08 ` Simmons, James A.
0 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-22 15:08 UTC (permalink / raw)
To: 'Dan Carpenter', James Simmons
Cc: devel@driverdev.osuosl.org, James Simmons, James Simmons,
Greg Kroah-Hartman, Linux Kernel Mailing List, Oleg Drokin,
HPDD-discuss@ml01.01.org, lustre-devel@lists.lustre.org
>This patch does a lot of stuff all at once and it is hard to review. It
>could easily be broken into patches which are easy to review.
I have more very large patches. With breaking them up that means you are
going to see hundreds of patches coming from me.
>Ok in this next section we move functions around and rename them but
>also introduce some bad changes in the new function.
Thanks for pointing out these bugs. We are still carrying these bugs in the
OpenSFS branch. Once you approve these changes I will sync up lib-socket.c
in the OpenSFS branch.
P.S
Does the 3rd patch look okay to you?
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
2015-05-22 15:08 ` Simmons, James A.
@ 2015-05-22 15:39 ` Dan Carpenter
-1 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 15:39 UTC (permalink / raw)
To: lustre-devel
On Fri, May 22, 2015 at 03:08:44PM +0000, Simmons, James A. wrote:
> >This patch does a lot of stuff all at once and it is hard to review. It
> >could easily be broken into patches which are easy to review.
>
> I have more very large patches. With breaking them up that means you are
> going to see hundreds of patches coming from me.
>
The problem here is that we have two upstreams which have diverged and
you're hopefully merging them into one upstream for the future.
Breaking things up into patches is a pain and it basically means you
have to do a lot of the work a second time. It sucks for you and I
totally understand that... :(
We regulary review 100+ patch series so that's not a big deal.
We still have to review a ball of code with it broken up or not broken
up but when it's broken up then I have scripts to strip out much of the
mechanical changes.
> >Ok in this next section we move functions around and rename them but
> >also introduce some bad changes in the new function.
>
> Thanks for pointing out these bugs. We are still carrying these bugs in the
> OpenSFS branch. Once you approve these changes I will sync up lib-socket.c
> in the OpenSFS branch.
>
> P.S
> Does the 3rd patch look okay to you?
Yeah. That looks fine.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 15:39 ` Dan Carpenter
0 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 15:39 UTC (permalink / raw)
To: Simmons, James A.
Cc: James Simmons, devel@driverdev.osuosl.org, James Simmons,
James Simmons, Greg Kroah-Hartman, Linux Kernel Mailing List,
Oleg Drokin, HPDD-discuss@ml01.01.org,
lustre-devel@lists.lustre.org
On Fri, May 22, 2015 at 03:08:44PM +0000, Simmons, James A. wrote:
> >This patch does a lot of stuff all at once and it is hard to review. It
> >could easily be broken into patches which are easy to review.
>
> I have more very large patches. With breaking them up that means you are
> going to see hundreds of patches coming from me.
>
The problem here is that we have two upstreams which have diverged and
you're hopefully merging them into one upstream for the future.
Breaking things up into patches is a pain and it basically means you
have to do a lot of the work a second time. It sucks for you and I
totally understand that... :(
We regulary review 100+ patch series so that's not a big deal.
We still have to review a ball of code with it broken up or not broken
up but when it's broken up then I have scripts to strip out much of the
mechanical changes.
> >Ok in this next section we move functions around and rename them but
> >also introduce some bad changes in the new function.
>
> Thanks for pointing out these bugs. We are still carrying these bugs in the
> OpenSFS branch. Once you approve these changes I will sync up lib-socket.c
> in the OpenSFS branch.
>
> P.S
> Does the 3rd patch look okay to you?
Yeah. That looks fine.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
2015-05-22 15:08 ` Simmons, James A.
@ 2015-05-22 21:40 ` Greg Kroah-Hartman
-1 siblings, 0 replies; 44+ messages in thread
From: Greg Kroah-Hartman @ 2015-05-22 21:40 UTC (permalink / raw)
To: lustre-devel
On Fri, May 22, 2015 at 03:08:44PM +0000, Simmons, James A. wrote:
> >This patch does a lot of stuff all at once and it is hard to review. It
> >could easily be broken into patches which are easy to review.
>
> I have more very large patches. With breaking them up that means you are
> going to see hundreds of patches coming from me.
hundreds of patches are not a problem at all for me, please do it that
way as what you sent is almost impossible to properly review.
thanks,
greg k-h
^ permalink raw reply [flat|nested] 44+ messages in thread
* Re: [lustre-devel] [PATCH 1/3] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 21:40 ` Greg Kroah-Hartman
0 siblings, 0 replies; 44+ messages in thread
From: Greg Kroah-Hartman @ 2015-05-22 21:40 UTC (permalink / raw)
To: Simmons, James A.
Cc: 'Dan Carpenter', James Simmons,
devel@driverdev.osuosl.org, James Simmons, James Simmons,
Linux Kernel Mailing List, Oleg Drokin, HPDD-discuss@ml01.01.org,
lustre-devel@lists.lustre.org
On Fri, May 22, 2015 at 03:08:44PM +0000, Simmons, James A. wrote:
> >This patch does a lot of stuff all at once and it is hard to review. It
> >could easily be broken into patches which are easy to review.
>
> I have more very large patches. With breaking them up that means you are
> going to see hundreds of patches coming from me.
hundreds of patches are not a problem at all for me, please do it that
way as what you sent is almost impossible to properly review.
thanks,
greg k-h
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: lustre-devel
From: James Simmons <uja.ornl@yahoo.com>
Currently the lnet headers used for user land applications
contain various kernel definations. This is due to the
fact libcfs contains kernel wrappers for user land which
will be going away. This patch sorted the header data
so all kernel containing structures are moved out of
headers that user land will use.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13792
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 4 -
.../lustre/include/linux/libcfs/libcfs_private.h | 51 --
.../lustre/include/linux/lnet/api-support.h | 44 --
drivers/staging/lustre/include/linux/lnet/api.h | 51 +-
.../staging/lustre/include/linux/lnet/lib-lnet.h | 249 +-------
.../staging/lustre/include/linux/lnet/lib-types.h | 649 +++++++++-----------
.../lustre/include/linux/lnet/linux/api-support.h | 42 --
.../lustre/include/linux/lnet/linux/lib-lnet.h | 71 ---
.../lustre/include/linux/lnet/linux/lib-types.h | 45 --
.../staging/lustre/include/linux/lnet/linux/lnet.h | 56 --
.../lustre/include/linux/lnet/lnet-sysctl.h | 49 --
drivers/staging/lustre/include/linux/lnet/lnet.h | 11 +-
.../staging/lustre/include/linux/lnet/lnetctl.h | 36 +-
drivers/staging/lustre/include/linux/lnet/lnetst.h | 665 ++++++++++++--------
drivers/staging/lustre/include/linux/lnet/nidstr.h | 73 +++
drivers/staging/lustre/include/linux/lnet/ptllnd.h | 93 ---
.../lustre/include/linux/lnet/ptllnd_wire.h | 119 ----
.../staging/lustre/include/linux/lnet/socklnd.h | 91 ++--
drivers/staging/lustre/include/linux/lnet/types.h | 266 +++++++--
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h | 9 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 1 -
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 4 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 2 -
drivers/staging/lustre/lnet/lnet/api-ni.c | 100 +---
drivers/staging/lustre/lnet/lnet/lib-eq.c | 2 +-
drivers/staging/lustre/lnet/lnet/lib-md.c | 6 +-
drivers/staging/lustre/lnet/lnet/lib-me.c | 24 +-
drivers/staging/lustre/lnet/lnet/lib-msg.c | 22 +-
drivers/staging/lustre/lnet/lnet/module.c | 4 +-
drivers/staging/lustre/lnet/lnet/router.c | 145 -----
drivers/staging/lustre/lustre/include/lustre_net.h | 4 +-
32 files changed, 1149 insertions(+), 1841 deletions(-)
delete mode 100644 drivers/staging/lustre/include/linux/lnet/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
create mode 100644 drivers/staging/lustre/include/linux/lnet/nidstr.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index f54bab6..aae9487 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -46,10 +46,6 @@
#include "curproc.h"
-#ifndef offsetof
-# define offsetof(typ, memb) ((long)(long_ptr_t)((char *)&(((typ *)0)->memb)))
-#endif
-
#ifndef ARRAY_SIZE
#define ARRAY_SIZE(a) ((sizeof(a)) / (sizeof((a)[0])))
#endif
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
index fef8825..6174946 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
@@ -410,36 +410,6 @@ int cfs_percpt_atomic_summary(atomic_t **refs);
*/
#define CLASSERT(cond) do {switch (42) {case (cond): case 0: break; } } while (0)
-/* support decl needed both by kernel and liblustre */
-int libcfs_isknown_lnd(int type);
-char *libcfs_lnd2modname(int type);
-char *libcfs_lnd2str(int type);
-int libcfs_str2lnd(const char *str);
-char *libcfs_net2str(__u32 net);
-char *libcfs_nid2str(lnet_nid_t nid);
-__u32 libcfs_str2net(const char *str);
-lnet_nid_t libcfs_str2nid(const char *str);
-int libcfs_str2anynid(lnet_nid_t *nid, const char *str);
-char *libcfs_id2str(lnet_process_id_t id);
-void cfs_free_nidlist(struct list_head *list);
-int cfs_parse_nidlist(char *str, int len, struct list_head *list);
-int cfs_match_nid(lnet_nid_t nid, struct list_head *list);
-
-/** \addtogroup lnet_addr
- * @{ */
-/* how an LNET NID encodes net:address */
-/** extract the address part of an lnet_nid_t */
-#define LNET_NIDADDR(nid) ((__u32)((nid) & 0xffffffff))
-/** extract the network part of an lnet_nid_t */
-#define LNET_NIDNET(nid) ((__u32)(((nid) >> 32)) & 0xffffffff)
-/** make an lnet_nid_t from a network part and an address part */
-#define LNET_MKNID(net, addr) ((((__u64)(net))<<32)|((__u64)(addr)))
-/* how net encodes type:number */
-#define LNET_NETNUM(net) ((net) & 0xffff)
-#define LNET_NETTYP(net) (((net) >> 16) & 0xffff)
-#define LNET_MKNET(typ, num) ((((__u32)(typ))<<16)|((__u32)(num)))
-/** @} lnet_addr */
-
/* max value for numeric network address */
#define MAX_NUMERIC_VALUE 0xffffffff
@@ -532,25 +502,4 @@ do { \
ptr += cfs_size_round(len + 1); \
} while (0)
-/**
- * Lustre Network Driver types.
- */
-enum {
- /* Only add to these values (i.e. don't ever change or redefine them):
- * network addresses depend on them... */
- QSWLND = 1,
- SOCKLND = 2,
- GMLND = 3, /* obsolete, keep it so that libcfs_nid2str works */
- PTLLND = 4,
- O2IBLND = 5,
- CIBLND = 6,
- OPENIBLND = 7,
- IIBLND = 8,
- LOLND = 9,
- RALND = 10,
- VIBLND = 11,
- MXLND = 12,
- GNILND = 13,
-};
-
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/api-support.h b/drivers/staging/lustre/include/linux/lnet/api-support.h
deleted file mode 100644
index 8f7fa28..0000000
--- a/drivers/staging/lustre/include/linux/lnet/api-support.h
+++ /dev/null
@@ -1,44 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_API_SUPPORT_H__
-#define __LNET_API_SUPPORT_H__
-
-#include "linux/api-support.h"
-
-#include "../libcfs/libcfs.h"
-#include "types.h"
-#include "lnet.h"
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/api.h b/drivers/staging/lustre/include/linux/lnet/api.h
index cd86517..76fb6fd 100644
--- a/drivers/staging/lustre/include/linux/lnet/api.h
+++ b/drivers/staging/lustre/include/linux/lnet/api.h
@@ -15,21 +15,19 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2011, 2015, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_API_H__
@@ -41,20 +39,15 @@
*
* LNet is an asynchronous message-passing API, which provides an unreliable
* connectionless service that can't guarantee any order. It supports OFA IB,
- * TCP/IP, and Cray Portals, and routes between heterogeneous networks.
- *
- * LNet can run both in OS kernel space and in userspace as a library.
+ * TCP/IP, and Cray Interconnects, and routes between heterogeneous networks.
* @{
*/
-#include "../lnet/types.h"
+#include "types.h"
/** \defgroup lnet_init_fini Initialization and cleanup
* The LNet must be properly initialized before any LNet calls can be made.
* @{ */
-int LNetInit(void);
-void LNetFini(void);
-
int LNetNIInit(lnet_pid_t requested_pid);
int LNetNIFini(void);
/** @} lnet_init_fini */
@@ -84,6 +77,7 @@ void LNetSnprintHandle(char *str, int str_len, lnet_handle_any_t handle);
/** @} lnet_addr */
+
/** \defgroup lnet_me Match entries
*
* A match entry (abbreviated as ME) describes a set of criteria to accept
@@ -98,16 +92,16 @@ void LNetSnprintHandle(char *str, int str_len, lnet_handle_any_t handle);
* @{ */
int LNetMEAttach(unsigned int portal,
lnet_process_id_t match_id_in,
- __u64 match_bits_in,
- __u64 ignore_bits_in,
+ __u64 match_bits_in,
+ __u64 ignore_bits_in,
lnet_unlink_t unlink_in,
lnet_ins_pos_t pos_in,
lnet_handle_me_t *handle_out);
int LNetMEInsert(lnet_handle_me_t current_in,
lnet_process_id_t match_id_in,
- __u64 match_bits_in,
- __u64 ignore_bits_in,
+ __u64 match_bits_in,
+ __u64 ignore_bits_in,
lnet_unlink_t unlink_in,
lnet_ins_pos_t position_in,
lnet_handle_me_t *handle_out);
@@ -128,11 +122,11 @@ int LNetMEUnlink(lnet_handle_me_t current_in);
* associated with a MD: LNetMDUnlink().
* @{ */
int LNetMDAttach(lnet_handle_me_t current_in,
- lnet_md_t md_in,
+ lnet_md_t md_in,
lnet_unlink_t unlink_in,
lnet_handle_md_t *handle_out);
-int LNetMDBind(lnet_md_t md_in,
+int LNetMDBind(lnet_md_t md_in,
lnet_unlink_t unlink_in,
lnet_handle_md_t *handle_out);
@@ -170,14 +164,15 @@ int LNetEQFree(lnet_handle_eq_t eventq_in);
int LNetEQGet(lnet_handle_eq_t eventq_in,
lnet_event_t *event_out);
+
int LNetEQWait(lnet_handle_eq_t eventq_in,
lnet_event_t *event_out);
int LNetEQPoll(lnet_handle_eq_t *eventqs_in,
- int neq_in,
- int timeout_ms,
+ int neq_in,
+ int timeout_ms,
lnet_event_t *event_out,
- int *which_eq_out);
+ int *which_eq_out);
/** @} lnet_eq */
/** \defgroup lnet_data Data movement operations
@@ -185,23 +180,24 @@ int LNetEQPoll(lnet_handle_eq_t *eventqs_in,
* The LNet API provides two data movement operations: LNetPut()
* and LNetGet().
* @{ */
-int LNetPut(lnet_nid_t self,
+int LNetPut(lnet_nid_t self,
lnet_handle_md_t md_in,
lnet_ack_req_t ack_req_in,
lnet_process_id_t target_in,
unsigned int portal_in,
- __u64 match_bits_in,
+ __u64 match_bits_in,
unsigned int offset_in,
- __u64 hdr_data_in);
+ __u64 hdr_data_in);
-int LNetGet(lnet_nid_t self,
+int LNetGet(lnet_nid_t self,
lnet_handle_md_t md_in,
lnet_process_id_t target_in,
unsigned int portal_in,
- __u64 match_bits_in,
+ __u64 match_bits_in,
unsigned int offset_in);
/** @} lnet_data */
+
/** \defgroup lnet_misc Miscellaneous operations.
* Miscellaneous operations.
* @{ */
@@ -209,7 +205,6 @@ int LNetGet(lnet_nid_t self,
int LNetSetLazyPortal(int portal);
int LNetClearLazyPortal(int portal);
int LNetCtl(unsigned int cmd, void *arg);
-int LNetSetAsync(lnet_process_id_t id, int nasync);
/** @} lnet_misc */
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index c00a2ca..5eacae4 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,56 +23,46 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2011, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lib-lnet.h
- *
- * Top level include for library side routines
*/
#ifndef __LNET_LIB_LNET_H__
#define __LNET_LIB_LNET_H__
-#include "linux/lib-lnet.h"
#include "../libcfs/libcfs.h"
-#include "types.h"
+#include "api.h"
#include "lnet.h"
#include "lib-types.h"
-extern lnet_t the_lnet; /* THE network */
-
-#if defined(LNET_USE_LIB_FREELIST)
-/* 1 CPT, simplify implementation... */
-# define LNET_CPT_MAX_BITS 0
-
-#else /* KERNEL and no freelist */
+extern lnet_t the_lnet; /* THE network */
-# if (BITS_PER_LONG == 32)
+#if (BITS_PER_LONG == 32)
/* 2 CPTs, allowing more CPTs might make us under memory pressure */
-# define LNET_CPT_MAX_BITS 1
+# define LNET_CPT_MAX_BITS 1
-# else /* 64-bit system */
+#else /* 64-bit system */
/*
* 256 CPTs for thousands of CPUs, allowing more CPTs might make us
* under risk of consuming all lh_cookie.
*/
-# define LNET_CPT_MAX_BITS 8
-# endif /* BITS_PER_LONG == 32 */
-#endif
+# define LNET_CPT_MAX_BITS 8
+#endif /* BITS_PER_LONG == 32 */
/* max allowed CPT number */
-#define LNET_CPT_MAX (1 << LNET_CPT_MAX_BITS)
+#define LNET_CPT_MAX (1 << LNET_CPT_MAX_BITS)
-#define LNET_CPT_NUMBER (the_lnet.ln_cpt_number)
-#define LNET_CPT_BITS (the_lnet.ln_cpt_bits)
-#define LNET_CPT_MASK ((1ULL << LNET_CPT_BITS) - 1)
+#define LNET_CPT_NUMBER (the_lnet.ln_cpt_number)
+#define LNET_CPT_BITS (the_lnet.ln_cpt_bits)
+#define LNET_CPT_MASK ((1ULL << LNET_CPT_BITS) - 1)
/** exclusive lock */
-#define LNET_LOCK_EX CFS_PERCPT_LOCK_EX
+#define LNET_LOCK_EX CFS_PERCPT_LOCK_EX
static inline int lnet_is_wire_handle_none(lnet_handle_wire_t *wh)
{
@@ -177,191 +163,9 @@ lnet_net_lock_current(void)
#define MAX_PORTALS 64
-/* these are only used by code with LNET_USE_LIB_FREELIST, but we still
- * exported them to !LNET_USE_LIB_FREELIST for easy implementation */
-#define LNET_FL_MAX_MES 2048
-#define LNET_FL_MAX_MDS 2048
-#define LNET_FL_MAX_EQS 512
-#define LNET_FL_MAX_MSGS 2048 /* Outstanding messages */
-
-#ifdef LNET_USE_LIB_FREELIST
-
-int lnet_freelist_init(lnet_freelist_t *fl, int n, int size);
-void lnet_freelist_fini(lnet_freelist_t *fl);
-
-static inline void *
-lnet_freelist_alloc(lnet_freelist_t *fl)
-{
- /* ALWAYS called with liblock held */
- lnet_freeobj_t *o;
-
- if (list_empty(&fl->fl_list))
- return NULL;
-
- o = list_entry(fl->fl_list.next, lnet_freeobj_t, fo_list);
- list_del(&o->fo_list);
- return (void *)&o->fo_contents;
-}
-
-static inline void
-lnet_freelist_free(lnet_freelist_t *fl, void *obj)
-{
- /* ALWAYS called with liblock held */
- lnet_freeobj_t *o = list_entry(obj, lnet_freeobj_t, fo_contents);
-
- list_add(&o->fo_list, &fl->fl_list);
-}
-
static inline lnet_eq_t *
lnet_eq_alloc(void)
{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = &the_lnet.ln_eq_container;
- lnet_eq_t *eq;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- eq = (lnet_eq_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- return eq;
-}
-
-static inline void
-lnet_eq_free_locked(lnet_eq_t *eq)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = &the_lnet.ln_eq_container;
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, eq);
-}
-
-static inline void
-lnet_eq_free(lnet_eq_t *eq)
-{
- lnet_res_lock(0);
- lnet_eq_free_locked(eq);
- lnet_res_unlock(0);
-}
-
-static inline lnet_libmd_t *
-lnet_md_alloc(lnet_md_t *umd)
-{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
- lnet_libmd_t *md;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- md = (lnet_libmd_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- if (md != NULL)
- INIT_LIST_HEAD(&md->md_list);
-
- return md;
-}
-
-static inline void
-lnet_md_free_locked(lnet_libmd_t *md)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, md);
-}
-
-static inline void
-lnet_md_free(lnet_libmd_t *md)
-{
- lnet_res_lock(0);
- lnet_md_free_locked(md);
- lnet_res_unlock(0);
-}
-
-static inline lnet_me_t *
-lnet_me_alloc(void)
-{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
- lnet_me_t *me;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- me = (lnet_me_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- return me;
-}
-
-static inline void
-lnet_me_free_locked(lnet_me_t *me)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, me);
-}
-
-static inline void
-lnet_me_free(lnet_me_t *me)
-{
- lnet_res_lock(0);
- lnet_me_free_locked(me);
- lnet_res_unlock(0);
-}
-
-static inline lnet_msg_t *
-lnet_msg_alloc(void)
-{
- /* NEVER called with network lock held */
- struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
- lnet_msg_t *msg;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_net_lock(0);
- msg = (lnet_msg_t *)lnet_freelist_alloc(&msc->msc_freelist);
- lnet_net_unlock(0);
-
- if (msg != NULL) {
- /* NULL pointers, clear flags etc */
- memset(msg, 0, sizeof(*msg));
- }
- return msg;
-}
-
-static inline void
-lnet_msg_free_locked(lnet_msg_t *msg)
-{
- /* ALWAYS called with network lock held */
- struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- LASSERT(!msg->msg_onactivelist);
- lnet_freelist_free(&msc->msc_freelist, msg);
-}
-
-static inline void
-lnet_msg_free(lnet_msg_t *msg)
-{
- lnet_net_lock(0);
- lnet_msg_free_locked(msg);
- lnet_net_unlock(0);
-}
-
-#else /* !LNET_USE_LIB_FREELIST */
-
-static inline lnet_eq_t *
-lnet_eq_alloc(void)
-{
- /* NEVER called with liblock held */
lnet_eq_t *eq;
LIBCFS_ALLOC(eq, sizeof(*eq));
@@ -371,14 +175,12 @@ lnet_eq_alloc(void)
static inline void
lnet_eq_free(lnet_eq_t *eq)
{
- /* ALWAYS called with resource lock held */
LIBCFS_FREE(eq, sizeof(*eq));
}
static inline lnet_libmd_t *
lnet_md_alloc(lnet_md_t *umd)
{
- /* NEVER called with liblock held */
lnet_libmd_t *md;
unsigned int size;
unsigned int niov;
@@ -407,7 +209,6 @@ lnet_md_alloc(lnet_md_t *umd)
static inline void
lnet_md_free(lnet_libmd_t *md)
{
- /* ALWAYS called with resource lock held */
unsigned int size;
if ((md->md_options & LNET_MD_KIOV) != 0)
@@ -421,7 +222,6 @@ lnet_md_free(lnet_libmd_t *md)
static inline lnet_me_t *
lnet_me_alloc(void)
{
- /* NEVER called with liblock held */
lnet_me_t *me;
LIBCFS_ALLOC(me, sizeof(*me));
@@ -431,14 +231,12 @@ lnet_me_alloc(void)
static inline void
lnet_me_free(lnet_me_t *me)
{
- /* ALWAYS called with resource lock held */
LIBCFS_FREE(me, sizeof(*me));
}
static inline lnet_msg_t *
lnet_msg_alloc(void)
{
- /* NEVER called with liblock held */
lnet_msg_t *msg;
LIBCFS_ALLOC(msg, sizeof(*msg));
@@ -450,18 +248,10 @@ lnet_msg_alloc(void)
static inline void
lnet_msg_free(lnet_msg_t *msg)
{
- /* ALWAYS called with network lock held */
LASSERT(!msg->msg_onactivelist);
LIBCFS_FREE(msg, sizeof(*msg));
}
-#define lnet_eq_free_locked(eq) lnet_eq_free(eq)
-#define lnet_md_free_locked(md) lnet_md_free(md)
-#define lnet_me_free_locked(me) lnet_me_free(me)
-#define lnet_msg_free_locked(msg) lnet_msg_free(msg)
-
-#endif /* LNET_USE_LIB_FREELIST */
-
lnet_libhandle_t *lnet_res_lh_lookup(struct lnet_res_container *rec,
__u64 cookie);
void lnet_res_lh_initialize(struct lnet_res_container *rec,
@@ -644,6 +434,9 @@ lnet_ni_t *lnet_nid2ni_locked(lnet_nid_t nid, int cpt);
lnet_ni_t *lnet_net2ni_locked(__u32 net, int cpt);
lnet_ni_t *lnet_net2ni(__u32 net);
+int lnet_init(void);
+void lnet_fini(void);
+
int lnet_notify(lnet_ni_t *ni, lnet_nid_t peer, int alive, unsigned long when);
void lnet_notify_locked(lnet_peer_t *lp, int notifylnd, int alive,
unsigned long when);
@@ -737,7 +530,9 @@ void lnet_recv(lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed,
unsigned int offset, unsigned int mlen, unsigned int rlen);
lnet_msg_t *lnet_create_reply_msg(lnet_ni_t *ni, lnet_msg_t *get_msg);
void lnet_set_reply_msg_len(lnet_ni_t *ni, lnet_msg_t *msg, unsigned int len);
+
void lnet_finalize(lnet_ni_t *ni, lnet_msg_t *msg, int rc);
+
void lnet_drop_delayed_msg_list(struct list_head *head, char *reason);
void lnet_recv_delayed_msg_list(struct list_head *head);
@@ -887,12 +682,12 @@ void lnet_peer_tables_destroy(void);
int lnet_peer_tables_create(void);
void lnet_debug_peer(lnet_nid_t nid);
-static inline void lnet_peer_set_alive(lnet_peer_t *lp)
+static inline void
+lnet_peer_set_alive(lnet_peer_t *lp)
{
lp->lp_last_alive = lp->lp_last_query = get_seconds();
if (!lp->lp_alive)
lnet_notify_locked(lp, 0, 1, lp->lp_last_alive);
}
-
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-types.h b/drivers/staging/lustre/include/linux/lnet/lib-types.h
index 5053766..87acf14 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-types.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-types.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,155 +23,42 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lib-types.h
- *
- * Types used by the library side routines that do not need to be
- * exposed to the user application
*/
#ifndef __LNET_LIB_TYPES_H__
#define __LNET_LIB_TYPES_H__
-#include "linux/lib-types.h"
-
-#include "../libcfs/libcfs.h"
-#include <linux/list.h>
-#include "types.h"
-
-#define WIRE_ATTR __attribute__((packed))
-
-/* Packed version of lnet_process_id_t to transfer via network */
-typedef struct {
- lnet_nid_t nid;
- lnet_pid_t pid; /* node id / process id */
-} WIRE_ATTR lnet_process_id_packed_t;
-
-/* The wire handle's interface cookie only matches one network interface in
- * one epoch (i.e. new cookie when the interface restarts or the node
- * reboots). The object cookie only matches one object on that interface
- * during that object's lifetime (i.e. no cookie re-use). */
-typedef struct {
- __u64 wh_interface_cookie;
- __u64 wh_object_cookie;
-} WIRE_ATTR lnet_handle_wire_t;
-
-typedef enum {
- LNET_MSG_ACK = 0,
- LNET_MSG_PUT,
- LNET_MSG_GET,
- LNET_MSG_REPLY,
- LNET_MSG_HELLO,
-} lnet_msg_type_t;
-
-/* The variant fields of the portals message header are aligned on an 8
- * byte boundary in the message header. Note that all types used in these
- * wire structs MUST be fixed size and the smaller types are placed at the
- * end. */
-typedef struct lnet_ack {
- lnet_handle_wire_t dst_wmd;
- __u64 match_bits;
- __u32 mlength;
-} WIRE_ATTR lnet_ack_t;
-
-typedef struct lnet_put {
- lnet_handle_wire_t ack_wmd;
- __u64 match_bits;
- __u64 hdr_data;
- __u32 ptl_index;
- __u32 offset;
-} WIRE_ATTR lnet_put_t;
-
-typedef struct lnet_get {
- lnet_handle_wire_t return_wmd;
- __u64 match_bits;
- __u32 ptl_index;
- __u32 src_offset;
- __u32 sink_length;
-} WIRE_ATTR lnet_get_t;
-
-typedef struct lnet_reply {
- lnet_handle_wire_t dst_wmd;
-} WIRE_ATTR lnet_reply_t;
-
-typedef struct lnet_hello {
- __u64 incarnation;
- __u32 type;
-} WIRE_ATTR lnet_hello_t;
+#include <linux/kthread.h>
+#include <linux/uio.h>
+#include <linux/types.h>
+#include <net/sock.h>
-typedef struct {
- lnet_nid_t dest_nid;
- lnet_nid_t src_nid;
- lnet_pid_t dest_pid;
- lnet_pid_t src_pid;
- __u32 type; /* lnet_msg_type_t */
- __u32 payload_length; /* payload data to follow */
- /*<------__u64 aligned------->*/
- union {
- lnet_ack_t ack;
- lnet_put_t put;
- lnet_get_t get;
- lnet_reply_t reply;
- lnet_hello_t hello;
- } msg;
-} WIRE_ATTR lnet_hdr_t;
-
-/* A HELLO message contains a magic number and protocol version
- * code in the header's dest_nid, the peer's NID in the src_nid, and
- * LNET_MSG_HELLO in the type field. All other common fields are zero
- * (including payload_size; i.e. no payload).
- * This is for use by byte-stream LNDs (e.g. TCP/IP) to check the peer is
- * running the same protocol and to find out its NID. These LNDs should
- * exchange HELLO messages when a connection is first established. Individual
- * LNDs can put whatever else they fancy in lnet_hdr_t::msg.
- */
-typedef struct {
- __u32 magic; /* LNET_PROTO_TCP_MAGIC */
- __u16 version_major; /* increment on incompatible change */
- __u16 version_minor; /* increment on compatible change */
-} WIRE_ATTR lnet_magicversion_t;
-
-/* PROTO MAGIC for LNDs */
-#define LNET_PROTO_IB_MAGIC 0x0be91b91
-#define LNET_PROTO_RA_MAGIC 0x0be91b92
-#define LNET_PROTO_QSW_MAGIC 0x0be91b93
-#define LNET_PROTO_GNI_MAGIC 0xb00fbabe /* ask Kim */
-#define LNET_PROTO_TCP_MAGIC 0xeebc0ded
-#define LNET_PROTO_PTL_MAGIC 0x50746C4E /* 'PtlN' unique magic */
-#define LNET_PROTO_MX_MAGIC 0x4d583130 /* 'MX10'! */
-#define LNET_PROTO_ACCEPTOR_MAGIC 0xacce7100
-#define LNET_PROTO_PING_MAGIC 0x70696E67 /* 'ping' */
-
-/* Placeholder for a future "unified" protocol across all LNDs */
-/* Current LNDs that receive a request with this magic will respond with a
- * "stub" reply using their current protocol */
-#define LNET_PROTO_MAGIC 0x45726963 /* ! */
-
-#define LNET_PROTO_TCP_VERSION_MAJOR 1
-#define LNET_PROTO_TCP_VERSION_MINOR 0
-
-/* Acceptor connection request */
-typedef struct {
- __u32 acr_magic; /* PTL_ACCEPTOR_PROTO_MAGIC */
- __u32 acr_version; /* protocol version */
- __u64 acr_nid; /* target NID */
-} WIRE_ATTR lnet_acceptor_connreq_t;
+#include "lnetctl.h"
-#define LNET_PROTO_ACCEPTOR_VERSION 1
+/* Max payload size */
+# define LNET_MAX_PAYLOAD CONFIG_LNET_MAX_PAYLOAD
+# if (LNET_MAX_PAYLOAD < LNET_MTU)
+# error "LNET_MAX_PAYLOAD too small - error in configure --with-max-payload-mb"
+# elif (LNET_MAX_PAYLOAD > (PAGE_SIZE * LNET_MAX_IOV))
+# error "LNET_MAX_PAYLOAD too large - error in configure --with-max-payload-mb"
+# endif
/* forward refs */
struct lnet_libmd;
typedef struct lnet_msg {
- struct list_head msg_activelist;
- struct list_head msg_list; /* Q for credits/MD */
+ struct list_head msg_activelist;
+ /* Q for credits/MD */
+ struct list_head msg_list;
- lnet_process_id_t msg_target;
+ lnet_process_id_t msg_target;
/* where is it from, it's only for building event */
lnet_nid_t msg_from;
__u32 msg_type;
@@ -195,110 +78,115 @@ typedef struct lnet_msg {
/* ready for pending on RX delay list */
unsigned int msg_rx_ready_delay:1;
- unsigned int msg_vmflush:1; /* VM trying to free memory */
- unsigned int msg_target_is_router:1; /* sending to a router */
- unsigned int msg_routing:1; /* being forwarded */
- unsigned int msg_ack:1; /* ack on finalize (PUT) */
- unsigned int msg_sending:1; /* outgoing message */
- unsigned int msg_receiving:1; /* being received */
- unsigned int msg_txcredit:1; /* taken an NI send credit */
- unsigned int msg_peertxcredit:1; /* taken a peer send credit */
- unsigned int msg_rtrcredit:1; /* taken a global router credit */
- unsigned int msg_peerrtrcredit:1; /* taken a peer router credit */
- unsigned int msg_onactivelist:1; /* on the activelist */
-
- struct lnet_peer *msg_txpeer; /* peer I'm sending to */
- struct lnet_peer *msg_rxpeer; /* peer I received from */
-
- void *msg_private;
- struct lnet_libmd *msg_md;
-
- unsigned int msg_len;
- unsigned int msg_wanted;
- unsigned int msg_offset;
- unsigned int msg_niov;
- struct kvec *msg_iov;
- lnet_kiov_t *msg_kiov;
-
- lnet_event_t msg_ev;
- lnet_hdr_t msg_hdr;
+ /* VM trying to free memory */
+ unsigned int msg_vmflush:1;
+ /* sending to a router */
+ unsigned int msg_target_is_router:1;
+ /* being forwarded */
+ unsigned int msg_routing:1;
+ /* ack on finalize (PUT) */
+ unsigned int msg_ack:1;
+ /* outgoing message */
+ unsigned int msg_sending:1;
+ /* being received */
+ unsigned int msg_receiving:1;
+ /* taken an NI send credit */
+ unsigned int msg_txcredit:1;
+ /* taken a peer send credit */
+ unsigned int msg_peertxcredit:1;
+ /* taken a global router credit */
+ unsigned int msg_rtrcredit:1;
+ /* taken a peer router credit */
+ unsigned int msg_peerrtrcredit:1;
+ /* on the activelist */
+ unsigned int msg_onactivelist:1;
+
+ /* peer I'm sending to */
+ struct lnet_peer *msg_txpeer;
+ /* peer I received from */
+ struct lnet_peer *msg_rxpeer;
+
+ void *msg_private;
+ struct lnet_libmd *msg_md;
+
+ unsigned int msg_len;
+ unsigned int msg_wanted;
+ unsigned int msg_offset;
+ unsigned int msg_niov;
+ struct kvec *msg_iov;
+ lnet_kiov_t *msg_kiov;
+
+ lnet_event_t msg_ev;
+ lnet_hdr_t msg_hdr;
} lnet_msg_t;
typedef struct lnet_libhandle {
- struct list_head lh_hash_chain;
- __u64 lh_cookie;
+ struct list_head lh_hash_chain;
+ __u64 lh_cookie;
} lnet_libhandle_t;
#define lh_entry(ptr, type, member) \
((type *)((char *)(ptr)-(char *)(&((type *)0)->member)))
typedef struct lnet_eq {
- struct list_head eq_list;
+ struct list_head eq_list;
lnet_libhandle_t eq_lh;
lnet_seq_t eq_enq_seq;
lnet_seq_t eq_deq_seq;
unsigned int eq_size;
lnet_eq_handler_t eq_callback;
lnet_event_t *eq_events;
- int **eq_refs; /* percpt refcount for EQ */
+ /* percpt refcount for EQ */
+ int **eq_refs;
} lnet_eq_t;
typedef struct lnet_me {
- struct list_head me_list;
- lnet_libhandle_t me_lh;
- lnet_process_id_t me_match_id;
- unsigned int me_portal;
- unsigned int me_pos; /* hash offset in mt_hash */
- __u64 me_match_bits;
- __u64 me_ignore_bits;
- lnet_unlink_t me_unlink;
- struct lnet_libmd *me_md;
+ struct list_head me_list;
+ lnet_libhandle_t me_lh;
+ lnet_process_id_t me_match_id;
+ unsigned int me_portal;
+ /* hash offset in mt_hash */
+ unsigned int me_pos;
+ __u64 me_match_bits;
+ __u64 me_ignore_bits;
+ lnet_unlink_t me_unlink;
+ struct lnet_libmd *me_md;
} lnet_me_t;
typedef struct lnet_libmd {
- struct list_head md_list;
- lnet_libhandle_t md_lh;
- lnet_me_t *md_me;
- char *md_start;
- unsigned int md_offset;
- unsigned int md_length;
- unsigned int md_max_size;
- int md_threshold;
- int md_refcount;
- unsigned int md_options;
- unsigned int md_flags;
- void *md_user_ptr;
- lnet_eq_t *md_eq;
- unsigned int md_niov; /* # frags */
+ struct list_head md_list;
+ lnet_libhandle_t md_lh;
+ lnet_me_t *md_me;
+ char *md_start;
+ unsigned int md_offset;
+ unsigned int md_length;
+ unsigned int md_max_size;
+ int md_threshold;
+ int md_refcount;
+ unsigned int md_options;
+ unsigned int md_flags;
+ void *md_user_ptr;
+ lnet_eq_t *md_eq;
+ /* # frags */
+ unsigned int md_niov;
union {
- struct kvec iov[LNET_MAX_IOV];
- lnet_kiov_t kiov[LNET_MAX_IOV];
+ struct kvec iov[LNET_MAX_IOV];
+ lnet_kiov_t kiov[LNET_MAX_IOV];
} md_iov;
} lnet_libmd_t;
-#define LNET_MD_FLAG_ZOMBIE (1 << 0)
-#define LNET_MD_FLAG_AUTO_UNLINK (1 << 1)
-#define LNET_MD_FLAG_ABORTED (1 << 2)
+#define LNET_MD_FLAG_ZOMBIE (1 << 0)
+#define LNET_MD_FLAG_AUTO_UNLINK (1 << 1)
+#define LNET_MD_FLAG_ABORTED (1 << 2)
-#ifdef LNET_USE_LIB_FREELIST
+/* info about peers we are trying to fail */
typedef struct {
- void *fl_objs; /* single contiguous array of objects */
- int fl_nobjs; /* the number of them */
- int fl_objsize; /* the size (including overhead) of each of them */
- struct list_head fl_list; /* where they are enqueued */
-} lnet_freelist_t;
-
-typedef struct {
- struct list_head fo_list; /* enqueue on fl_list */
- void *fo_contents; /* aligned contents */
-} lnet_freeobj_t;
-#endif
-
-typedef struct {
- /* info about peers we are trying to fail */
- struct list_head tp_list; /* ln_test_peers */
- lnet_nid_t tp_nid; /* matching nid */
- unsigned int tp_threshold; /* # failures to simulate */
+ /* ln_test_peers */
+ struct list_head tp_list;
+ /* matching nid */
+ lnet_nid_t tp_nid;
+ /* # failures to simulate */
+ unsigned int tp_threshold;
} lnet_test_peer_t;
#define LNET_COOKIE_TYPE_MD 1
@@ -307,15 +195,17 @@ typedef struct {
#define LNET_COOKIE_TYPE_BITS 2
#define LNET_COOKIE_MASK ((1ULL << LNET_COOKIE_TYPE_BITS) - 1ULL)
-struct lnet_ni; /* forward ref */
+struct lnet_ni; /* forward ref */
+/* fields managed by portals */
typedef struct lnet_lnd {
- /* fields managed by portals */
- struct list_head lnd_list; /* stash in the LND table */
- int lnd_refcount; /* # active instances */
+ /* stash in the LND table */
+ struct list_head lnd_list;
+ /* # active instances */
+ int lnd_refcount;
/* fields initialised by the LND */
- unsigned int lnd_type;
+ __u32 lnd_type;
int (*lnd_startup)(struct lnet_ni *ni);
void (*lnd_shutdown)(struct lnet_ni *ni);
@@ -366,47 +256,57 @@ typedef struct lnet_lnd {
/* accept a new connection */
int (*lnd_accept)(struct lnet_ni *ni, struct socket *sock);
-
} lnd_t;
-#define LNET_NI_STATUS_UP 0x15aac0de
-#define LNET_NI_STATUS_DOWN 0xdeadface
-#define LNET_NI_STATUS_INVALID 0x00000000
typedef struct {
- lnet_nid_t ns_nid;
- __u32 ns_status;
- __u32 ns_unused;
+ lnet_nid_t ns_nid;
+ __u32 ns_status;
+ __u32 ns_unused;
} WIRE_ATTR lnet_ni_status_t;
struct lnet_tx_queue {
- int tq_credits; /* # tx credits free */
- int tq_credits_min; /* lowest it's been */
- int tq_credits_max; /* total # tx credits */
- struct list_head tq_delayed; /* delayed TXs */
+ /* # tx credits free */
+ int tq_credits;
+ /* lowest it's been */
+ int tq_credits_min;
+ /* total # tx credits */
+ int tq_credits_max;
+ /* delayed TXs */
+ struct list_head tq_delayed;
};
-#define LNET_MAX_INTERFACES 16
-
typedef struct lnet_ni {
spinlock_t ni_lock;
- struct list_head ni_list; /* chain on ln_nis */
- struct list_head ni_cptlist; /* chain on ln_nis_cpt */
- int ni_maxtxcredits; /* # tx credits */
+ /* chain on ln_nis */
+ struct list_head ni_list;
+ /* chain on ln_nis_cpt */
+ struct list_head ni_cptlist;
+ /* # tx credits */
+ int ni_maxtxcredits;
/* # per-peer send credits */
int ni_peertxcredits;
/* # per-peer router buffer credits */
int ni_peerrtrcredits;
/* seconds to consider peer dead */
int ni_peertimeout;
- int ni_ncpts; /* number of CPTs */
- __u32 *ni_cpts; /* bond NI on some CPTs */
- lnet_nid_t ni_nid; /* interface's NID */
- void *ni_data; /* instance-specific data */
- lnd_t *ni_lnd; /* procedural interface */
- struct lnet_tx_queue **ni_tx_queues; /* percpt TX queues */
- int **ni_refs; /* percpt reference count */
- long ni_last_alive; /* when I was last alive */
- lnet_ni_status_t *ni_status; /* my health status */
+ /* number of CPTs */
+ int ni_ncpts;
+ /* bond NI on some CPTs */
+ __u32 *ni_cpts;
+ /* interface's NID */
+ lnet_nid_t ni_nid;
+ /* instance-specific data */
+ void *ni_data;
+ /* procedural interface */
+ lnd_t *ni_lnd;
+ /* percpt TX queues */
+ struct lnet_tx_queue **ni_tx_queues;
+ /* percpt reference count */
+ int **ni_refs;
+ /* when I was last alive */
+ long ni_last_alive;
+ /* my health status */
+ lnet_ni_status_t *ni_status;
/* equivalent interfaces to use */
char *ni_interfaces[LNET_MAX_INTERFACES];
} lnet_ni_t;
@@ -435,43 +335,72 @@ typedef struct {
#define LNET_PINGINFO_SIZE offsetof(lnet_ping_info_t, pi_ni[LNET_MAX_RTR_NIS])
typedef struct {
/* chain on the_lnet.ln_zombie_rcd or ln_deathrow_rcd */
- struct list_head rcd_list;
- lnet_handle_md_t rcd_mdh; /* ping buffer MD */
- struct lnet_peer *rcd_gateway; /* reference to gateway */
- lnet_ping_info_t *rcd_pinginfo; /* ping buffer */
+ struct list_head rcd_list;
+ /* ping buffer MD */
+ lnet_handle_md_t rcd_mdh;
+ /* reference to gateway */
+ struct lnet_peer *rcd_gateway;
+ /* ping buffer */
+ lnet_ping_info_t *rcd_pinginfo;
} lnet_rc_data_t;
typedef struct lnet_peer {
- struct list_head lp_hashlist; /* chain on peer hash */
- struct list_head lp_txq; /* messages blocking for tx credits */
- struct list_head lp_rtrq; /* messages blocking for router credits */
- struct list_head lp_rtr_list; /* chain on router list */
- int lp_txcredits; /* # tx credits available */
- int lp_mintxcredits; /* low water mark */
- int lp_rtrcredits; /* # router credits */
- int lp_minrtrcredits; /* low water mark */
- unsigned int lp_alive:1; /* alive/dead? */
- unsigned int lp_notify:1; /* notification outstanding? */
- unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
- unsigned int lp_notifying:1; /* some thread is handling notification */
- unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
- int lp_alive_count; /* # times router went dead<->alive */
- long lp_txqnob; /* bytes queued for sending */
- unsigned long lp_timestamp; /* time of last aliveness news */
- unsigned long lp_ping_timestamp; /* time of last ping attempt */
- unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
- unsigned long lp_last_alive; /* when I was last alive */
- unsigned long lp_last_query; /* when lp_ni was queried last time */
- lnet_ni_t *lp_ni; /* interface peer is on */
- lnet_nid_t lp_nid; /* peer's NID */
- int lp_refcount; /* # refs */
- int lp_cpt; /* CPT this peer attached on */
+ /* chain on peer hash */
+ struct list_head lp_hashlist;
+ /* messages blocking for tx credits */
+ struct list_head lp_txq;
+ /* messages blocking for router credits */
+ struct list_head lp_rtrq;
+ /* chain on router list */
+ struct list_head lp_rtr_list;
+ /* # tx credits available */
+ int lp_txcredits;
+ /* low water mark */
+ int lp_mintxcredits;
+ /* # router credits */
+ int lp_rtrcredits;
+ /* low water mark */
+ int lp_minrtrcredits;
+ /* alive/dead? */
+ unsigned int lp_alive:1;
+ /* notification outstanding? */
+ unsigned int lp_notify:1;
+ /* outstanding notification for LND? */
+ unsigned int lp_notifylnd:1;
+ /* some thread is handling notification */
+ unsigned int lp_notifying:1;
+ /* SEND event outstanding from ping */
+ unsigned int lp_ping_notsent;
+ /* # times router went dead<->alive */
+ int lp_alive_count;
+ /* bytes queued for sending */
+ long lp_txqnob;
+ /* time of last aliveness news */
+ unsigned long lp_timestamp;
+ /* time of last ping attempt */
+ unsigned long lp_ping_timestamp;
+ /* != 0 if ping reply expected */
+ unsigned long lp_ping_deadline;
+ /* when I was last alive */
+ unsigned long lp_last_alive;
+ /* when lp_ni was queried last time */
+ unsigned long lp_last_query;
+ /* interface peer is on */
+ lnet_ni_t *lp_ni;
+ /* peer's NID */
+ lnet_nid_t lp_nid;
+ /* # refs */
+ int lp_refcount;
+ /* CPT this peer attached on */
+ int lp_cpt;
/* # refs from lnet_route_t::lr_gateway */
int lp_rtr_refcount;
/* returned RC ping features */
unsigned int lp_ping_feats;
- struct list_head lp_routes; /* routers on this peer */
- lnet_rc_data_t *lp_rcd; /* router checker state */
+ /* routers on this peer */
+ struct list_head lp_routes;
+ /* router checker state */
+ lnet_rc_data_t *lp_rcd;
} lnet_peer_t;
/* peer hash size */
@@ -480,10 +409,14 @@ typedef struct lnet_peer {
/* peer hash table */
struct lnet_peer_table {
- int pt_version; /* /proc validity stamp */
- int pt_number; /* # peers extant */
- struct list_head pt_deathrow; /* zombie peers */
- struct list_head *pt_hash; /* NID->peer hash */
+ /* /proc validity stamp */
+ int pt_version;
+ /* # peers extant */
+ int pt_number;
+ /* zombie peers */
+ struct list_head pt_deathrow;
+ /* NID->peer hash */
+ struct list_head *pt_hash;
};
/* peer aliveness is enabled only on routers for peers in a network where the
@@ -492,14 +425,22 @@ struct lnet_peer_table {
(lp)->lp_ni->ni_peertimeout > 0)
typedef struct {
- struct list_head lr_list; /* chain on net */
- struct list_head lr_gwlist; /* chain on gateway */
- lnet_peer_t *lr_gateway; /* router node */
- __u32 lr_net; /* remote network number */
- int lr_seq; /* sequence for round-robin */
- unsigned int lr_downis; /* number of down NIs */
- unsigned int lr_hops; /* how far I am */
- unsigned int lr_priority; /* route priority */
+ /* chain on net */
+ struct list_head lr_list;
+ /* chain on gateway */
+ struct list_head lr_gwlist;
+ /* router node */
+ lnet_peer_t *lr_gateway;
+ /* remote network number */
+ __u32 lr_net;
+ /* sequence for round-robin */
+ int lr_seq;
+ /* number of down NIs */
+ unsigned int lr_downis;
+ /* how far I am */
+ unsigned int lr_hops;
+ /* route priority */
+ unsigned int lr_priority;
} lnet_route_t;
#define LNET_REMOTE_NETS_HASH_DEFAULT (1U << 7)
@@ -507,43 +448,41 @@ typedef struct {
#define LNET_REMOTE_NETS_HASH_SIZE (1 << the_lnet.ln_remote_nets_hbits)
typedef struct {
- struct list_head lrn_list; /* chain on ln_remote_nets_hash */
- struct list_head lrn_routes; /* routes to me */
- __u32 lrn_net; /* my net number */
+ /* chain on ln_remote_nets_hash */
+ struct list_head lrn_list;
+ /* routes to me */
+ struct list_head lrn_routes;
+ /* my net number */
+ __u32 lrn_net;
} lnet_remotenet_t;
typedef struct {
- struct list_head rbp_bufs; /* my free buffer pool */
- struct list_head rbp_msgs; /* messages blocking for a buffer */
- int rbp_npages; /* # pages in each buffer */
- int rbp_nbuffers; /* # buffers */
- int rbp_credits; /* # free buffers / blocked messages */
- int rbp_mincredits; /* low water mark */
+ /* my free buffer pool */
+ struct list_head rbp_bufs;
+ /* messages blocking for a buffer */
+ struct list_head rbp_msgs;
+ /* # pages in each buffer */
+ int rbp_npages;
+ /* # buffers */
+ int rbp_nbuffers;
+ /* # free buffers / blocked messages */
+ int rbp_credits;
+ /* low water mark */
+ int rbp_mincredits;
} lnet_rtrbufpool_t;
typedef struct {
- struct list_head rb_list; /* chain on rbp_bufs */
- lnet_rtrbufpool_t *rb_pool; /* owning pool */
- lnet_kiov_t rb_kiov[0]; /* the buffer space */
+ /* chain on rbp_bufs */
+ struct list_head rb_list;
+ /* owning pool */
+ lnet_rtrbufpool_t *rb_pool;
+ /* the buffer space */
+ lnet_kiov_t rb_kiov[0];
} lnet_rtrbuf_t;
-typedef struct {
- __u32 msgs_alloc;
- __u32 msgs_max;
- __u32 errors;
- __u32 send_count;
- __u32 recv_count;
- __u32 route_count;
- __u32 drop_count;
- __u64 send_length;
- __u64 recv_length;
- __u64 route_length;
- __u64 drop_length;
-} WIRE_ATTR lnet_counters_t;
-
-#define LNET_PEER_HASHSIZE 503 /* prime! */
-
-#define LNET_NRBPOOLS 3 /* # different router buffer pools */
+#define LNET_PEER_HASHSIZE 503 /* prime! */
+
+#define LNET_NRBPOOLS 3 /* # different router buffer pools */
enum {
/* Didn't match anything */
@@ -559,9 +498,9 @@ enum {
};
/* Options for lnet_portal_t::ptl_options */
-#define LNET_PTL_LAZY (1 << 0)
-#define LNET_PTL_MATCH_UNIQUE (1 << 1) /* unique match, for RDMA */
-#define LNET_PTL_MATCH_WILDCARD (1 << 2) /* wildcard match, request portal */
+#define LNET_PTL_LAZY (1 << 0)
+#define LNET_PTL_MATCH_UNIQUE (1 << 1) /* unique match, for RDMA */
+#define LNET_PTL_MATCH_WILDCARD (1 << 2) /* wildcard match, request portal */
/* parameter for matching operations (GET, PUT) */
struct lnet_match_info {
@@ -591,13 +530,15 @@ struct lnet_match_info {
struct lnet_match_table {
/* reserved for upcoming patches, CPU partition ID */
unsigned int mt_cpt;
- unsigned int mt_portal; /* portal index */
+ /* portal index */
+ unsigned int mt_portal;
/* match table is set as "enabled" if there's non-exhausted MD
* attached on mt_mhash, it's only valid for wildcard portal */
unsigned int mt_enabled;
/* bitmap to flag whether MEs on mt_hash are exhausted or not */
__u64 mt_exhausted[LNET_MT_EXHAUSTED_BMAP];
- struct list_head *mt_mhash; /* matching hash */
+ /* matching hash */
+ struct list_head *mt_mhash;
};
/* these are only useful for wildcard portal */
@@ -612,21 +553,22 @@ struct lnet_match_table {
typedef struct lnet_portal {
spinlock_t ptl_lock;
- unsigned int ptl_index; /* portal ID, reserved */
+ /* portal ID, reserved */
+ unsigned int ptl_index;
/* flags on this portal: lazy, unique... */
unsigned int ptl_options;
/* list of messages which are stealing buffer */
- struct list_head ptl_msg_stealing;
+ struct list_head ptl_msg_stealing;
/* messages blocking for MD */
- struct list_head ptl_msg_delayed;
+ struct list_head ptl_msg_delayed;
/* Match table for each CPT */
struct lnet_match_table **ptl_mtables;
/* spread rotor of incoming "PUT" */
unsigned int ptl_rotor;
/* # active entries for this portal */
- int ptl_mt_nmaps;
+ int ptl_mt_nmaps;
/* array of active entries' cpu-partition-id */
- int ptl_mt_maps[0];
+ int ptl_mt_maps[0];
} lnet_portal_t;
#define LNET_LH_HASH_BITS 12
@@ -635,28 +577,28 @@ typedef struct lnet_portal {
/* resource container (ME, MD, EQ) */
struct lnet_res_container {
- unsigned int rec_type; /* container type */
- __u64 rec_lh_cookie; /* cookie generator */
- struct list_head rec_active; /* active resource list */
- struct list_head *rec_lh_hash; /* handle hash */
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_t rec_freelist; /* freelist for resources */
-#endif
+ /* container type */
+ unsigned int rec_type;
+ /* cookie generator */
+ __u64 rec_lh_cookie;
+ /* active resource list */
+ struct list_head rec_active;
+ /* handle hash */
+ struct list_head *rec_lh_hash;
};
/* message container */
struct lnet_msg_container {
- int msc_init; /* initialized or not */
+ /* initialized or not */
+ int msc_init;
/* max # threads finalizing */
int msc_nfinalizers;
/* msgs waiting to complete finalizing */
- struct list_head msc_finalizing;
- struct list_head msc_active; /* active message list */
+ struct list_head msc_finalizing;
+ /* active message list */
+ struct list_head msc_active;
/* threads doing finalization */
void **msc_finalizers;
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_t msc_freelist; /* freelist for messages */
-#endif
};
/* Router Checker states */
@@ -684,7 +626,7 @@ typedef struct {
/* Event Queue container */
struct lnet_res_container ln_eq_container;
- wait_queue_head_t ln_eq_waitq;
+ wait_queue_head_t ln_eq_waitq;
spinlock_t ln_eq_wait_lock;
unsigned int ln_remote_nets_hbits;
@@ -695,23 +637,25 @@ typedef struct {
lnet_counters_t **ln_counters;
struct lnet_peer_table **ln_peer_tables;
/* failure simulation */
- struct list_head ln_test_peers;
+ struct list_head ln_test_peers;
- struct list_head ln_nis; /* LND instances */
+ /* LND instances */
+ struct list_head ln_nis;
/* NIs bond on specific CPT(s) */
- struct list_head ln_nis_cpt;
+ struct list_head ln_nis_cpt;
/* dying LND instances */
- struct list_head ln_nis_zombie;
- lnet_ni_t *ln_loni; /* the loopback NI */
+ struct list_head ln_nis_zombie;
+ /* the loopback NI */
+ lnet_ni_t *ln_loni;
/* NI to wait for events in */
lnet_ni_t *ln_eq_waitni;
/* remote networks with routes to them */
- struct list_head *ln_remote_nets_hash;
+ struct list_head *ln_remote_nets_hash;
/* validity stamp */
__u64 ln_remote_nets_version;
/* list of all known routers */
- struct list_head ln_routers;
+ struct list_head ln_routers;
/* validity stamp */
__u64 ln_routers_version;
/* percpt router buffer pools */
@@ -726,15 +670,16 @@ typedef struct {
/* router checker's event queue */
lnet_handle_eq_t ln_rc_eqh;
/* rcd still pending on net */
- struct list_head ln_rcd_deathrow;
+ struct list_head ln_rcd_deathrow;
/* rcd ready for free */
- struct list_head ln_rcd_zombie;
+ struct list_head ln_rcd_zombie;
/* serialise startup/shutdown */
struct semaphore ln_rc_signal;
struct mutex ln_api_mutex;
struct mutex ln_lnd_mutex;
- int ln_init; /* LNetInit() called? */
+ /* LNetInit() called? */
+ int ln_init;
/* Have I called LNetNIInit myself? */
int ln_niinit_self;
/* LNetNIInit/LNetNIFini counter */
@@ -742,12 +687,14 @@ typedef struct {
/* shutdown in progress */
int ln_shutdown;
- int ln_routing; /* am I a router? */
- lnet_pid_t ln_pid; /* requested pid */
+ /* am I a router? */
+ int ln_routing;
+ /* requested pid */
+ lnet_pid_t ln_pid;
/* uniquely identifies this ni in this epoch */
__u64 ln_interface_cookie;
/* registered LNDs */
- struct list_head ln_lnds;
+ struct list_head ln_lnds;
/* space for network names */
char *ln_network_tokens;
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/api-support.h b/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
deleted file mode 100644
index e237ad6..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
+++ /dev/null
@@ -1,42 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LINUX_API_SUPPORT_H__
-#define __LINUX_API_SUPPORT_H__
-
-#ifndef __LNET_API_SUPPORT_H__
-#error Do not #include this file directly. #include <lnet /api-support.h> instead
-#endif
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
deleted file mode 100644
index 0f8f04d..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
+++ /dev/null
@@ -1,71 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LIB_LNET_H__
-#define __LNET_LINUX_LIB_LNET_H__
-
-#ifndef __LNET_LIB_LNET_H__
-#error Do not #include this file directly. #include <linux/lnet/lib-lnet.h> instead
-#endif
-
-# include <asm/page.h>
-# include <linux/string.h>
-# include <asm/io.h>
-#include "../../libcfs/libcfs.h"
-
-static inline __u64
-lnet_page2phys(struct page *p)
-{
- /* compiler optimizer will elide unused branches */
-
- switch (sizeof(typeof(page_to_phys(p)))) {
- case 4:
- /* page_to_phys returns a 32 bit physical address. This must
- * be a 32 bit machine with <= 4G memory and we must ensure we
- * don't sign extend when converting to 64 bits. */
- return (unsigned long)page_to_phys(p);
-
- case 8:
- /* page_to_phys returns a 64 bit physical address :) */
- return page_to_phys(p);
-
- default:
- LBUG();
- return 0;
- }
-}
-
-#define LNET_ROUTER
-
-#endif /* __LNET_LINUX_LIB_LNET_H__ */
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
deleted file mode 100644
index 669e8c0..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
+++ /dev/null
@@ -1,45 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LIB_TYPES_H__
-#define __LNET_LINUX_LIB_TYPES_H__
-
-#ifndef __LNET_LIB_TYPES_H__
-#error Do not #include this file directly. #include <linux/lnet/lib-types.h> instead
-#endif
-
-# include <linux/uio.h>
-# include <linux/types.h>
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
deleted file mode 100644
index 1e888f1..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
+++ /dev/null
@@ -1,56 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2011, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LNET_H__
-#define __LNET_LINUX_LNET_H__
-
-#ifndef __LNET_H__
-#error Do not #include this file directly. #include <linux/lnet/lnet.h> instead
-#endif
-
-/*
- * lnet.h
- *
- * User application interface file
- */
-
-#include <linux/uio.h>
-#include <linux/types.h>
-
-#define cfs_tcp_sendpage(sk, page, offset, size, flags) \
- tcp_sendpage(sk, page, offset, size, flags)
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h b/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
deleted file mode 100644
index 2dee1b9..0000000
--- a/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
+++ /dev/null
@@ -1,49 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_SYSCTL_H__
-#define __LNET_SYSCTL_H__
-
-#if defined(CONFIG_SYSCTL)
-
-#define CTL_KRANAL 201
-#define CTL_O2IBLND 205
-#define CTL_PTLLND 206
-#define CTL_QSWNAL 207
-#define CTL_SOCKLND 208
-#define CTL_GNILND 210
-
-#endif
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet.h b/drivers/staging/lustre/include/linux/lnet/lnet.h
index 75c0ab9..a31b29c 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnet.h
@@ -26,10 +26,12 @@
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_H__
@@ -40,12 +42,7 @@
*
* User application interface file
*/
-#include "linux/lnet.h"
-
#include "types.h"
-#include "api.h"
-
-#define LNET_NIDSTR_COUNT 1024 /* # of nidstrings */
-#define LNET_NIDSTR_SIZE 32 /* size of each one (see below for usage) */
+#include "nidstr.h"
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetctl.h b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
index 98181d3..2b82a06 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnetctl.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
@@ -14,27 +14,29 @@
* along with Portals; if not, write to the Free Software
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*
- * header for libptlctl.a
+ * header for lnet ioctl
*/
-#ifndef _PTLCTL_H_
-#define _PTLCTL_H_
+/*
+ * Copyright (c) 2014, Intel Corporation.
+ */
+#ifndef _LNETCTL_H_
+#define _LNETCTL_H_
-#include "../libcfs/libcfs.h"
#include "types.h"
-#define LNET_DEV_ID 0
-#define LNET_DEV_PATH "/dev/lnet"
-#define LNET_DEV_MAJOR 10
-#define LNET_DEV_MINOR 240
-#define OBD_DEV_ID 1
-#define OBD_DEV_NAME "obd"
-#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME
-#define OBD_DEV_MAJOR 10
-#define OBD_DEV_MINOR 241
-#define SMFS_DEV_ID 2
-#define SMFS_DEV_PATH "/dev/snapdev"
-#define SMFS_DEV_MAJOR 10
-#define SMFS_DEV_MINOR 242
+#define LNET_DEV_ID 0
+#define LNET_DEV_PATH "/dev/lnet"
+#define LNET_DEV_MAJOR 10
+#define LNET_DEV_MINOR 240
+#define OBD_DEV_ID 1
+#define OBD_DEV_NAME "obd"
+#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME
+#define OBD_DEV_MAJOR 10
+#define OBD_DEV_MINOR 241
+#define SMFS_DEV_ID 2
+#define SMFS_DEV_PATH "/dev/snapdev"
+#define SMFS_DEV_MAJOR 10
+#define SMFS_DEV_MINOR 242
int ptl_initialize(int argc, char **argv);
int jt_ptl_network(int argc, char **argv);
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetst.h b/drivers/staging/lustre/include/linux/lnet/lnetst.h
index 885f708..462eae8 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnetst.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnetst.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,123 +23,182 @@
* Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, Intel Corporation.
+ * Copyright (c) 2012, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lnetst.h
*
- * Author: Liang Zhen <liangzhen@clusterfs.com>
+ * Author: Liang Zhen <liang.zhen@intel.com>
*/
#ifndef __LNET_ST_H__
#define __LNET_ST_H__
-#include "../libcfs/libcfs.h"
-#include "lnet.h"
-#include "lib-types.h"
-
#define LST_FEAT_NONE (0)
#define LST_FEAT_BULK_LEN (1 << 0) /* enable variable page size */
#define LST_FEATS_EMPTY (LST_FEAT_NONE)
#define LST_FEATS_MASK (LST_FEAT_NONE | LST_FEAT_BULK_LEN)
-#define LST_NAME_SIZE 32 /* max name buffer length */
-
-#define LSTIO_DEBUG 0xC00 /* debug */
-#define LSTIO_SESSION_NEW 0xC01 /* create session */
-#define LSTIO_SESSION_END 0xC02 /* end session */
-#define LSTIO_SESSION_INFO 0xC03 /* query session */
-#define LSTIO_GROUP_ADD 0xC10 /* add group */
-#define LSTIO_GROUP_LIST 0xC11 /* list all groups in session */
-#define LSTIO_GROUP_INFO 0xC12 /* query default information of specified group */
-#define LSTIO_GROUP_DEL 0xC13 /* delete group */
-#define LSTIO_NODES_ADD 0xC14 /* add nodes to specified group */
-#define LSTIO_GROUP_UPDATE 0xC15 /* update group */
-#define LSTIO_BATCH_ADD 0xC20 /* add batch */
-#define LSTIO_BATCH_START 0xC21 /* start batch */
-#define LSTIO_BATCH_STOP 0xC22 /* stop batch */
-#define LSTIO_BATCH_DEL 0xC23 /* delete batch */
-#define LSTIO_BATCH_LIST 0xC24 /* show all batches in the session */
-#define LSTIO_BATCH_INFO 0xC25 /* show defail of specified batch */
-#define LSTIO_TEST_ADD 0xC26 /* add test (to batch) */
-#define LSTIO_BATCH_QUERY 0xC27 /* query batch status */
-#define LSTIO_STAT_QUERY 0xC30 /* get stats */
-
-typedef struct {
- lnet_nid_t ses_nid; /* nid of console node */
- __u64 ses_stamp; /* time stamp */
-} lst_sid_t; /*** session id */
+#define LST_NAME_SIZE 32 /* max name buffer length */
+
+/* debug */
+#define LSTIO_DEBUG 0xC00
+/* create session */
+#define LSTIO_SESSION_NEW 0xC01
+/* end session */
+#define LSTIO_SESSION_END 0xC02
+/* query session */
+#define LSTIO_SESSION_INFO 0xC03
+/* add group */
+#define LSTIO_GROUP_ADD 0xC10
+/* list all groups in session */
+#define LSTIO_GROUP_LIST 0xC11
+/* query default information of specified group */
+#define LSTIO_GROUP_INFO 0xC12
+/* delete group */
+#define LSTIO_GROUP_DEL 0xC13
+/* add nodes to specified group */
+#define LSTIO_NODES_ADD 0xC14
+/* update group */
+#define LSTIO_GROUP_UPDATE 0xC15
+/* add batch */
+#define LSTIO_BATCH_ADD 0xC20
+/* start batch */
+#define LSTIO_BATCH_START 0xC21
+/* stop batch */
+#define LSTIO_BATCH_STOP 0xC22
+/* delete batch */
+#define LSTIO_BATCH_DEL 0xC23
+/* show all batches in the session */
+#define LSTIO_BATCH_LIST 0xC24
+/* show details of specified batch */
+#define LSTIO_BATCH_INFO 0xC25
+/* add test (to batch) */
+#define LSTIO_TEST_ADD 0xC26
+/* query batch status */
+#define LSTIO_BATCH_QUERY 0xC27
+/* get stats */
+#define LSTIO_STAT_QUERY 0xC30
+
+/*** session id */
+typedef struct {
+ /* nid of console node */
+ lnet_nid_t ses_nid;
+ /* time stamp */
+ __u64 ses_stamp;
+} lst_sid_t;
extern lst_sid_t LST_INVALID_SID;
-typedef struct {
- __u64 bat_id; /* unique id in session */
-} lst_bid_t; /*** batch id (group of tests) */
-
-/* Status of test node */
-#define LST_NODE_ACTIVE 0x1 /* node in this session */
-#define LST_NODE_BUSY 0x2 /* node is taken by other session */
-#define LST_NODE_DOWN 0x4 /* node is down */
-#define LST_NODE_UNKNOWN 0x8 /* node not in session */
-
-typedef struct {
- lnet_process_id_t nde_id; /* id of node */
- int nde_state; /* state of node */
-} lstcon_node_ent_t; /*** node entry, for list_group command */
-
-typedef struct {
- int nle_nnode; /* # of nodes */
- int nle_nactive; /* # of active nodes */
- int nle_nbusy; /* # of busy nodes */
- int nle_ndown; /* # of down nodes */
- int nle_nunknown; /* # of unknown nodes */
-} lstcon_ndlist_ent_t; /*** node_list entry, for list_batch command */
-
-typedef struct {
- int tse_type; /* test type */
- int tse_loop; /* loop count */
- int tse_concur; /* concurrency of test */
-} lstcon_test_ent_t; /*** test summary entry, for list_batch command */
-
-typedef struct {
- int bae_state; /* batch status */
- int bae_timeout; /* batch timeout */
- int bae_ntest; /* # of tests in the batch */
-} lstcon_batch_ent_t; /*** batch summary entry, for list_batch command */
-
-typedef struct {
- lstcon_ndlist_ent_t tbe_cli_nle; /* client (group) node_list entry */
- lstcon_ndlist_ent_t tbe_srv_nle; /* server (group) node_list entry */
+/*** batch id (group of tests) */
+typedef struct {
+ /* unique id in session */
+ __u64 bat_id;
+} lst_bid_t;
+
+/*** Status of test node */
+
+/* node in this session */
+#define LST_NODE_ACTIVE 0x1
+/* node is taken by other session */
+#define LST_NODE_BUSY 0x2
+/* node is down */
+#define LST_NODE_DOWN 0x4
+/* node not in session */
+#define LST_NODE_UNKNOWN 0x8
+
+/*** node entry, for list_group command */
+typedef struct {
+ /* id of node */
+ lnet_process_id_t nde_id;
+ /* state of node */
+ int nde_state;
+} lstcon_node_ent_t;
+
+/*** node_list entry, for list_batch command */
+typedef struct {
+ /* # of nodes */
+ int nle_nnode;
+ /* # of active nodes */
+ int nle_nactive;
+ /* # of busy nodes */
+ int nle_nbusy;
+ /* # of down nodes */
+ int nle_ndown;
+ /* # of unknown nodes */
+ int nle_nunknown;
+} lstcon_ndlist_ent_t;
+
+/*** test summary entry, for list_batch command */
+typedef struct {
+ /* test type */
+ int tse_type;
+ /* loop count */
+ int tse_loop;
+ /* concurrency of test */
+ int tse_concur;
+} lstcon_test_ent_t;
+
+/*** batch summary entry, for list_batch command */
+typedef struct {
+ /* batch status */
+ int bae_state;
+ /* batch status */
+ int bae_timeout;
+ /* # of tests in the batch */
+ int bae_ntest;
+} lstcon_batch_ent_t;
+
+/*** test/batch verbose information entry, for list_batch command */
+typedef struct {
+ /* client (group) node_list entry */
+ lstcon_ndlist_ent_t tbe_cli_nle;
+ /* server (group) node_list entry */
+ lstcon_ndlist_ent_t tbe_srv_nle;
union {
- lstcon_test_ent_t tbe_test; /* test entry */
- lstcon_batch_ent_t tbe_batch; /* batch entry */
+ /* test entry */
+ lstcon_test_ent_t tbe_test;
+ /* batch entry */
+ lstcon_batch_ent_t tbe_batch;
} u;
-} lstcon_test_batch_ent_t; /*** test/batch verbose information entry,
- *** for list_batch command */
-
-typedef struct {
- struct list_head rpe_link; /* link chain */
- lnet_process_id_t rpe_peer; /* peer's id */
- struct timeval rpe_stamp; /* time stamp of RPC */
- int rpe_state; /* peer's state */
- int rpe_rpc_errno; /* RPC errno */
-
- lst_sid_t rpe_sid; /* peer's session id */
- int rpe_fwk_errno; /* framework errno */
- int rpe_priv[4]; /* private data */
- char rpe_payload[0]; /* private reply payload */
+} lstcon_test_batch_ent_t;
+
+typedef struct {
+ /* link chain */
+ struct list_head rpe_link;
+ /* peer's id */
+ lnet_process_id_t rpe_peer;
+ /* time stamp of RPC */
+ struct timeval rpe_stamp;
+ /* peer's state */
+ int rpe_state;
+ /* RPC errno */
+ int rpe_rpc_errno;
+ /* peer's session id */
+ lst_sid_t rpe_sid;
+ /* framework errno */
+ int rpe_fwk_errno;
+ /* private data */
+ int rpe_priv[4];
+ /* private reply payload */
+ char rpe_payload[0];
} lstcon_rpc_ent_t;
typedef struct {
- int trs_rpc_stat[4]; /* RPCs stat (0: total, 1: failed, 2: finished, 4: reserved */
- int trs_rpc_errno; /* RPC errno */
- int trs_fwk_stat[8]; /* framework stat */
- int trs_fwk_errno; /* errno of the first remote error */
- void *trs_fwk_private; /* private framework stat */
+ /* RPCs stat (0: total, 1: failed, 2: finished, 4: reserved */
+ int trs_rpc_stat[4];
+ /* RPC errno */
+ int trs_rpc_errno;
+ /* framework stat */
+ int trs_fwk_stat[8];
+ /* errno of the first remote error */
+ int trs_fwk_errno;
+ /* private framework stat */
+ void *trs_fwk_private;
} lstcon_trans_stat_t;
static inline int
@@ -238,237 +293,351 @@ lstcon_statqry_stat_failure(lstcon_trans_stat_t *stat, int inc)
/* create a session */
typedef struct {
- int lstio_ses_key; /* IN: local key */
- int lstio_ses_timeout; /* IN: session timeout */
- int lstio_ses_force; /* IN: force create ? */
+ /* IN: local key */
+ int lstio_ses_key;
+ /* IN: session timeout */
+ int lstio_ses_timeout;
+ /* IN: force create ? */
+ int lstio_ses_force;
/** IN: session features */
- unsigned lstio_ses_feats;
- lst_sid_t *lstio_ses_idp; /* OUT: session id */
- int lstio_ses_nmlen; /* IN: name length */
- char *lstio_ses_namep; /* IN: session name */
+ unsigned lstio_ses_feats;
+ /* OUT: session id */
+ lst_sid_t *lstio_ses_idp;
+ /* IN: name length */
+ int lstio_ses_nmlen;
+ /* IN: session name */
+ char *lstio_ses_namep;
} lstio_session_new_args_t;
/* query current session */
typedef struct {
- lst_sid_t *lstio_ses_idp; /* OUT: session id */
- int *lstio_ses_keyp; /* OUT: local key */
+ /* OUT: session id */
+ lst_sid_t *lstio_ses_idp;
+ /* OUT: local key */
+ int *lstio_ses_keyp;
/** OUT: session features */
- unsigned *lstio_ses_featp;
- lstcon_ndlist_ent_t *lstio_ses_ndinfo; /* OUT: */
- int lstio_ses_nmlen; /* IN: name length */
- char *lstio_ses_namep; /* OUT: session name */
+ unsigned *lstio_ses_featp;
+ /* OUT: */
+ lstcon_ndlist_ent_t *lstio_ses_ndinfo;
+ /* IN: name length */
+ int lstio_ses_nmlen;
+ /* OUT: session name */
+ char *lstio_ses_namep;
} lstio_session_info_args_t;
/* delete a session */
typedef struct {
- int lstio_ses_key; /* IN: session key */
+ /* IN: session key */
+ int lstio_ses_key;
} lstio_session_end_args_t;
-#define LST_OPC_SESSION 1
-#define LST_OPC_GROUP 2
-#define LST_OPC_NODES 3
+#define LST_OPC_SESSION 1
+#define LST_OPC_GROUP 2
+#define LST_OPC_NODES 3
#define LST_OPC_BATCHCLI 4
#define LST_OPC_BATCHSRV 5
typedef struct {
- int lstio_dbg_key; /* IN: session key */
- int lstio_dbg_type; /* IN: debug sessin|batch|group|nodes list */
- int lstio_dbg_flags; /* IN: reserved debug flags */
- int lstio_dbg_timeout; /* IN: timeout of debug */
-
- int lstio_dbg_nmlen; /* IN: len of name */
- char *lstio_dbg_namep; /* IN: name of group|batch */
- int lstio_dbg_count; /* IN: # of test nodes to debug */
- lnet_process_id_t *lstio_dbg_idsp; /* IN: id of test nodes */
- struct list_head *lstio_dbg_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_dbg_key;
+ /* IN: debug sessin|batch|group|nodes list */
+ int lstio_dbg_type;
+ /* IN: reserved debug flags */
+ int lstio_dbg_flags;
+ /* IN: timeout of debug */
+ int lstio_dbg_timeout;
+ /* IN: len of name */
+ int lstio_dbg_nmlen;
+ /* IN: name of group|batch */
+ char *lstio_dbg_namep;
+ /* IN: # of test nodes to debug */
+ int lstio_dbg_count;
+ /* IN: id of test nodes */
+ lnet_process_id_t *lstio_dbg_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_dbg_resultp;
} lstio_debug_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
} lstio_group_add_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen; /* IN: name length */
+ /* IN: group name */
+ char *lstio_grp_namep;
} lstio_group_del_args_t;
-#define LST_GROUP_CLEAN 1 /* remove inactive nodes in the group */
-#define LST_GROUP_REFRESH 2 /* refresh inactive nodes in the group */
-#define LST_GROUP_RMND 3 /* delete nodes from the group */
-
-typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_opc; /* IN: OPC */
- int lstio_grp_args; /* IN: arguments */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
- int lstio_grp_count; /* IN: # of nodes id */
- lnet_process_id_t *lstio_grp_idsp; /* IN: array of nodes */
- struct list_head *lstio_grp_resultp; /* OUT: list head of result buffer */
+#define LST_GROUP_CLEAN 1 /* remove inactive nodes in the group */
+#define LST_GROUP_REFRESH 2 /* refresh inactive nodes in the group */
+#define LST_GROUP_RMND 3 /* delete nodes from the group */
+
+typedef struct {
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: OPC */
+ int lstio_grp_opc;
+ /* IN: arguments */
+ int lstio_grp_args;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
+ /* IN: # of nodes id */
+ int lstio_grp_count;
+ /* IN: array of nodes */
+ lnet_process_id_t *lstio_grp_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_grp_resultp;
} lstio_group_update_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
- int lstio_grp_count; /* IN: # of nodes */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
+ /* IN: # of nodes */
+ int lstio_grp_count;
/** OUT: session features */
- unsigned *lstio_grp_featp;
- lnet_process_id_t *lstio_grp_idsp; /* IN: nodes */
- struct list_head *lstio_grp_resultp; /* OUT: list head of result buffer */
+ unsigned *lstio_grp_featp;
+ /* IN: nodes */
+ lnet_process_id_t *lstio_grp_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_grp_resultp;
} lstio_group_nodes_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_idx; /* IN: group idx */
- int lstio_grp_nmlen; /* IN: name len */
- char *lstio_grp_namep; /* OUT: name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: group idx */
+ int lstio_grp_idx;
+ /* IN: name len */
+ int lstio_grp_nmlen;
+ /* OUT: name */
+ char *lstio_grp_namep;
} lstio_group_list_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name len */
- char *lstio_grp_namep; /* IN: name */
- lstcon_ndlist_ent_t *lstio_grp_entp; /* OUT: description of group */
-
- int *lstio_grp_idxp; /* IN/OUT: node index */
- int *lstio_grp_ndentp; /* IN/OUT: # of nodent */
- lstcon_node_ent_t *lstio_grp_dentsp; /* OUT: nodent array */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name len */
+ int lstio_grp_nmlen;
+ /* IN: name */
+ char *lstio_grp_namep;
+ /* OUT: description of group */
+ lstcon_ndlist_ent_t *lstio_grp_entp;
+ /* IN/OUT: node index */
+ int *lstio_grp_idxp;
+ /* IN/OUT: # of nodent */
+ int *lstio_grp_ndentp;
+ /* OUT: nodent array */
+ lstcon_node_ent_t *lstio_grp_dentsp;
} lstio_group_info_args_t;
-#define LST_DEFAULT_BATCH "batch" /* default batch name */
+#define LST_DEFAULT_BATCH "batch" /* default batch name */
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_add_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key; /* IN: session key */
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_del_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_timeout; /* IN: timeout for the batch */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: timeout for the batch */
+ int lstio_bat_timeout;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_run_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_force; /* IN: abort unfinished test RPC */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: abort unfinished test RPC */
+ int lstio_bat_force;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_stop_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_testidx; /* IN: test index */
- int lstio_bat_client; /* IN: is test client? */
- int lstio_bat_timeout; /* IN: timeout for waiting */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: test index */
+ int lstio_bat_testidx;
+ /* IN: is test client? */
+ int lstio_bat_client;
+ /* IN: timeout for waiting */
+ int lstio_bat_timeout;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_query_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_idx; /* IN: index */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: index */
+ int lstio_bat_idx;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_list_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: name */
- int lstio_bat_server; /* IN: query server or not */
- int lstio_bat_testidx; /* IN: test index */
- lstcon_test_batch_ent_t *lstio_bat_entp; /* OUT: batch ent */
-
- int *lstio_bat_idxp; /* IN/OUT: index of node */
- int *lstio_bat_ndentp; /* IN/OUT: # of nodent */
- lstcon_node_ent_t *lstio_bat_dentsp; /* array of nodent */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: name */
+ char *lstio_bat_namep;
+ /* IN: query server or not */
+ int lstio_bat_server;
+ /* IN: test index */
+ int lstio_bat_testidx;
+ /* OUT: batch ent */
+ lstcon_test_batch_ent_t *lstio_bat_entp;
+ /* IN/OUT: index of node */
+ int *lstio_bat_idxp;
+ /* IN/OUT: # of nodent */
+ int *lstio_bat_ndentp;
+ /* array of nodent */
+ lstcon_node_ent_t *lstio_bat_dentsp;
} lstio_batch_info_args_t;
/* add stat in session */
typedef struct {
- int lstio_sta_key; /* IN: session key */
- int lstio_sta_timeout; /* IN: timeout for stat request */
- int lstio_sta_nmlen; /* IN: group name length */
- char *lstio_sta_namep; /* IN: group name */
- int lstio_sta_count; /* IN: # of pid */
- lnet_process_id_t *lstio_sta_idsp; /* IN: pid */
- struct list_head *lstio_sta_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_sta_key;
+ /* IN: timeout for stat request */
+ int lstio_sta_timeout;
+ /* IN: group name length */
+ int lstio_sta_nmlen;
+ /* IN: group name */
+ char *lstio_sta_namep;
+ /* IN: # of pid */
+ int lstio_sta_count;
+ /* IN: pid */
+ lnet_process_id_t *lstio_sta_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_sta_resultp;
} lstio_stat_args_t;
typedef enum {
- LST_TEST_BULK = 1,
- LST_TEST_PING = 2
+ LST_TEST_BULK = 1,
+ LST_TEST_PING = 2
} lst_test_type_t;
/* create a test in a batch */
-#define LST_MAX_CONCUR 1024 /* Max concurrency of test */
-
-typedef struct {
- int lstio_tes_key; /* IN: session key */
- int lstio_tes_bat_nmlen; /* IN: batch name len */
- char *lstio_tes_bat_name; /* IN: batch name */
- int lstio_tes_type; /* IN: test type */
- int lstio_tes_oneside; /* IN: one sided test */
- int lstio_tes_loop; /* IN: loop count */
- int lstio_tes_concur; /* IN: concurrency */
-
- int lstio_tes_dist; /* IN: node distribution in destination groups */
- int lstio_tes_span; /* IN: node span in destination groups */
- int lstio_tes_sgrp_nmlen; /* IN: source group name length */
- char *lstio_tes_sgrp_name; /* IN: group name */
- int lstio_tes_dgrp_nmlen; /* IN: destination group name length */
- char *lstio_tes_dgrp_name; /* IN: group name */
-
- int lstio_tes_param_len; /* IN: param buffer len */
- void *lstio_tes_param; /* IN: parameter for specified test:
- lstio_bulk_param_t,
- lstio_ping_param_t,
- ... more */
- int *lstio_tes_retp; /* OUT: private returned value */
- struct list_head *lstio_tes_resultp; /* OUT: list head of result buffer */
+#define LST_MAX_CONCUR 1024 /* Max concurrency of test */
+
+typedef struct {
+ /* IN: session key */
+ int lstio_tes_key;
+ /* IN: batch name len */
+ int lstio_tes_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_tes_bat_name;
+ /* IN: test type */
+ int lstio_tes_type;
+ /* IN: one sided test */
+ int lstio_tes_oneside;
+ /* IN: loop count */
+ int lstio_tes_loop;
+ /* IN: concurrency */
+ int lstio_tes_concur;
+ /* IN: node distribution in destination groups */
+ int lstio_tes_dist;
+ /* IN: node span in destination groups */
+ int lstio_tes_span;
+ /* IN: source group name length */
+ int lstio_tes_sgrp_nmlen;
+ /* IN: group name */
+ char *lstio_tes_sgrp_name;
+ /* IN: destination group name length */
+ int lstio_tes_dgrp_nmlen;
+ /* IN: group name */
+ char *lstio_tes_dgrp_name;
+ /* IN: param buffer len */
+ int lstio_tes_param_len;
+ /* IN: parameter for specified test:
+ * lstio_bulk_param_t,
+ * lstio_ping_param_t,
+ * ... more */
+ void *lstio_tes_param;
+ /* OUT: private returned value */
+ int *lstio_tes_retp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_tes_resultp;
} lstio_test_args_t;
typedef enum {
- LST_BRW_READ = 1,
- LST_BRW_WRITE = 2
+ LST_BRW_READ = 1,
+ LST_BRW_WRITE = 2
} lst_brw_type_t;
typedef enum {
- LST_BRW_CHECK_NONE = 1,
- LST_BRW_CHECK_SIMPLE = 2,
- LST_BRW_CHECK_FULL = 3
+ LST_BRW_CHECK_NONE = 1,
+ LST_BRW_CHECK_SIMPLE = 2,
+ LST_BRW_CHECK_FULL = 3
} lst_brw_flags_t;
typedef struct {
- int blk_opc; /* bulk operation code */
- int blk_size; /* size (bytes) */
- int blk_time; /* time of running the test*/
- int blk_flags; /* reserved flags */
+ /* bulk operation code */
+ int blk_opc;
+ /* size (bytes) */
+ int blk_size; /* size (bytes) */
+ /* time of running the test*/
+ int blk_time;
+ /* reserved flags */
+ int blk_flags;
} lst_test_bulk_param_t;
typedef struct {
- int png_size; /* size of ping message */
- int png_time; /* time */
- int png_loop; /* loop */
- int png_flags; /* reserved flags */
+ /* size of ping message */
+ int png_size;
+ /* time */
+ int png_time;
+ /* loop */
+ int png_loop;
+ /* reserved flags */
+ int png_flags;
} lst_test_ping_param_t;
-/* more tests */
typedef struct {
__u32 errors;
__u32 rpcs_sent;
diff --git a/drivers/staging/lustre/include/linux/lnet/nidstr.h b/drivers/staging/lustre/include/linux/lnet/nidstr.h
new file mode 100644
index 0000000..14cca80
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/nidstr.h
@@ -0,0 +1,73 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, 2014, Intel Corporation.
+ */
+#ifndef _LNET_NIDSTRINGS_H
+#define _LNET_NIDSTRINGS_H
+#include "types.h"
+
+/**
+ * Lustre Network Driver types.
+ */
+enum {
+ /* Only add to these values (i.e. don't ever change or redefine them):
+ * network addresses depend on them... */
+ QSWLND = 1,
+ SOCKLND = 2,
+ GMLND = 3,
+ PTLLND = 4,
+ O2IBLND = 5,
+ CIBLND = 6,
+ OPENIBLND = 7,
+ IIBLND = 8,
+ LOLND = 9,
+ RALND = 10,
+ VIBLND = 11,
+ MXLND = 12,
+ GNILND = 13,
+ GNIIPLND = 14,
+};
+
+/* # of nidstrings */
+#define LNET_NIDSTR_COUNT 1024
+/* size of each one (see below for usage) */
+#define LNET_NIDSTR_SIZE 32
+
+int libcfs_isknown_lnd(int type);
+char *libcfs_lnd2modname(int type);
+char *libcfs_lnd2str(int type);
+int libcfs_str2lnd(const char *str);
+char *libcfs_net2str(__u32 net);
+char *libcfs_nid2str(lnet_nid_t nid);
+__u32 libcfs_str2net(const char *str);
+lnet_nid_t libcfs_str2nid(const char *str);
+int libcfs_str2anynid(lnet_nid_t *nid, const char *str);
+char *libcfs_id2str(lnet_process_id_t id);
+void cfs_free_nidlist(struct list_head *list);
+int cfs_parse_nidlist(char *str, int len, struct list_head *list);
+int cfs_match_nid(lnet_nid_t nid, struct list_head *list);
+
+#endif /* _LNET_NIDSTRINGS_H */
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd.h b/drivers/staging/lustre/include/linux/lnet/ptllnd.h
deleted file mode 100644
index c91d653..0000000
--- a/drivers/staging/lustre/include/linux/lnet/ptllnd.h
+++ /dev/null
@@ -1,93 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lnet/include/lnet/ptllnd.h
- *
- * Author: PJ Kirner <pjkirner@clusterfs.com>
- */
-
-/*
- * The PTLLND was designed to support Portals with
- * Lustre and non-lustre UNLINK semantics.
- * However for now the two targets are Cray Portals
- * on the XT3 and Lustre Portals (for testing) both
- * have Lustre UNLINK semantics, so this is defined
- * by default.
- */
-#define LUSTRE_PORTALS_UNLINK_SEMANTICS
-
-#ifdef _USING_LUSTRE_PORTALS_
-
-/* NIDs are 64-bits on Lustre Portals */
-#define FMT_NID "%llu"
-#define FMT_PID "%d"
-
-/* When using Lustre Portals Lustre completion semantics are imlicit*/
-#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS 0
-
-#else /* _USING_CRAY_PORTALS_ */
-
-/* NIDs are integers on Cray Portals */
-#define FMT_NID "%u"
-#define FMT_PID "%d"
-
-/* When using Cray Portals this is defined in the Cray Portals Header*/
-/*#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS */
-
-/* Can compare handles directly on Cray Portals */
-#define PtlHandleIsEqual(a, b) ((a) == (b))
-
-/* Different error types on Cray Portals*/
-#define ptl_err_t ptl_ni_fail_t
-
-/*
- * The Cray Portals has no maximum number of IOVs. The
- * maximum is limited only by memory and size of the
- * int parameters (2^31-1).
- * Lustre only really require that the underyling
- * implementation to support at least LNET_MAX_IOV,
- * so for Cray portals we can safely just use that
- * value here.
- *
- */
-#define PTL_MD_MAX_IOV LNET_MAX_IOV
-
-#endif
-
-#define FMT_PTLID "ptlid:"FMT_PID"-"FMT_NID
-
-/* Align incoming small request messages to an 8 byte boundary if this is
- * supported to avoid alignment issues on some architectures */
-#ifndef PTL_MD_LOCAL_ALIGN8
-# define PTL_MD_LOCAL_ALIGN8 0
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h b/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
deleted file mode 100644
index 808f37b..0000000
--- a/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
+++ /dev/null
@@ -1,119 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lnet/include/lnet/ptllnd_wire.h
- *
- * Author: PJ Kirner <pjkirner@clusterfs.com>
- */
-
-/* Minimum buffer size that any peer will post to receive ptllnd messages */
-#define PTLLND_MIN_BUFFER_SIZE 256
-
-/************************************************************************
- * Tunable defaults that {u,k}lnds/ptllnd should have in common.
- */
-
-#define PTLLND_PORTAL 9 /* The same portal PTLPRC used when talking to cray portals */
-#define PTLLND_PID 9 /* The Portals PID */
-#define PTLLND_PEERCREDITS 8 /* concurrent sends to 1 peer */
-
-/* Default buffer size for kernel ptllnds (guaranteed eager) */
-#define PTLLND_MAX_KLND_MSG_SIZE 512
-
-/* Default buffer size for catamount ptllnds (not guaranteed eager) - large
- * enough to avoid RDMA for anything sent while control is not in liblustre */
-#define PTLLND_MAX_ULND_MSG_SIZE 512
-
-/************************************************************************
- * Portals LND Wire message format.
- * These are sent in sender's byte order (i.e. receiver flips).
- */
-
-#define PTL_RESERVED_MATCHBITS 0x100 /* below this value is reserved
- * above is for bulk data transfer */
-#define LNET_MSG_MATCHBITS 0 /* the value for the message channel */
-
-typedef struct {
- lnet_hdr_t kptlim_hdr; /* portals header */
- char kptlim_payload[0]; /* piggy-backed payload */
-} WIRE_ATTR kptl_immediate_msg_t;
-
-typedef struct {
- lnet_hdr_t kptlrm_hdr; /* portals header */
- __u64 kptlrm_matchbits; /* matchbits */
-} WIRE_ATTR kptl_rdma_msg_t;
-
-typedef struct {
- __u64 kptlhm_matchbits; /* matchbits */
- __u32 kptlhm_max_msg_size; /* max message size */
-} WIRE_ATTR kptl_hello_msg_t;
-
-typedef struct {
- /* First 2 fields fixed FOR ALL TIME */
- __u32 ptlm_magic; /* I'm a Portals LND message */
- __u16 ptlm_version; /* this is my version number */
- __u8 ptlm_type; /* the message type */
- __u8 ptlm_credits; /* returned credits */
- __u32 ptlm_nob; /* # bytes in whole message */
- __u32 ptlm_cksum; /* checksum (0 == no checksum) */
- __u64 ptlm_srcnid; /* sender's NID */
- __u64 ptlm_srcstamp; /* sender's incarnation */
- __u64 ptlm_dstnid; /* destination's NID */
- __u64 ptlm_dststamp; /* destination's incarnation */
- __u32 ptlm_srcpid; /* sender's PID */
- __u32 ptlm_dstpid; /* destination's PID */
-
- union {
- kptl_immediate_msg_t immediate;
- kptl_rdma_msg_t rdma;
- kptl_hello_msg_t hello;
- } WIRE_ATTR ptlm_u;
-
-} kptl_msg_t;
-
-/* kptl_msg_t::ptlm_credits is only a __u8 */
-#define PTLLND_MSG_MAX_CREDITS ((typeof(((kptl_msg_t *)0)->ptlm_credits)) - 1)
-
-#define PTLLND_MSG_MAGIC LNET_PROTO_PTL_MAGIC
-#define PTLLND_MSG_VERSION 0x04
-
-#define PTLLND_RDMA_OK 0x00
-#define PTLLND_RDMA_FAIL 0x01
-
-#define PTLLND_MSG_TYPE_INVALID 0x00
-#define PTLLND_MSG_TYPE_PUT 0x01
-#define PTLLND_MSG_TYPE_GET 0x02
-#define PTLLND_MSG_TYPE_IMMEDIATE 0x03 /* No bulk data xfer*/
-#define PTLLND_MSG_TYPE_NOOP 0x04
-#define PTLLND_MSG_TYPE_HELLO 0x05
-#define PTLLND_MSG_TYPE_NAK 0x06
diff --git a/drivers/staging/lustre/include/linux/lnet/socklnd.h b/drivers/staging/lustre/include/linux/lnet/socklnd.h
index 389038b..e62df8c 100644
--- a/drivers/staging/lustre/include/linux/lnet/socklnd.h
+++ b/drivers/staging/lustre/include/linux/lnet/socklnd.h
@@ -15,89 +15,92 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2012, 2014 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/socklnd.h
- *
- * #defines shared between socknal implementation and utilities
*/
#ifndef __LNET_LNET_SOCKLND_H__
#define __LNET_LNET_SOCKLND_H__
#include "types.h"
-#include "lib-types.h"
-#define SOCKLND_CONN_NONE (-1)
+#define SOCKLND_CONN_NONE (-1)
#define SOCKLND_CONN_ANY 0
-#define SOCKLND_CONN_CONTROL 1
-#define SOCKLND_CONN_BULK_IN 2
-#define SOCKLND_CONN_BULK_OUT 3
-#define SOCKLND_CONN_NTYPES 4
+#define SOCKLND_CONN_CONTROL 1
+#define SOCKLND_CONN_BULK_IN 2
+#define SOCKLND_CONN_BULK_OUT 3
+#define SOCKLND_CONN_NTYPES 4
#define SOCKLND_CONN_ACK SOCKLND_CONN_BULK_IN
typedef struct {
- __u32 kshm_magic; /* magic number of socklnd message */
- __u32 kshm_version; /* version of socklnd message */
- lnet_nid_t kshm_src_nid; /* sender's nid */
- lnet_nid_t kshm_dst_nid; /* destination nid */
- lnet_pid_t kshm_src_pid; /* sender's pid */
- lnet_pid_t kshm_dst_pid; /* destination pid */
- __u64 kshm_src_incarnation; /* sender's incarnation */
- __u64 kshm_dst_incarnation; /* destination's incarnation */
- __u32 kshm_ctype; /* connection type */
- __u32 kshm_nips; /* # IP addrs */
- __u32 kshm_ips[0]; /* IP addrs */
+ /* magic number of socklnd message */
+ __u32 kshm_magic;
+ /* version of socklnd message */
+ __u32 kshm_version;
+ /* sender's nid */
+ lnet_nid_t kshm_src_nid;
+ /* destination nid */
+ lnet_nid_t kshm_dst_nid;
+ /* sender's pid */
+ lnet_pid_t kshm_src_pid;
+ /* destination pid */
+ lnet_pid_t kshm_dst_pid;
+ /* sender's incarnation */
+ __u64 kshm_src_incarnation;
+ /* destination's incarnation */
+ __u64 kshm_dst_incarnation;
+ /* connection type */
+ __u32 kshm_ctype;
+ /* # IP addrs */
+ __u32 kshm_nips;
+ __u32 kshm_ips[0];
} WIRE_ATTR ksock_hello_msg_t;
typedef struct {
- lnet_hdr_t ksnm_hdr; /* lnet hdr */
-
- /*
- * ksnm_payload is removed because of winnt compiler's limitation:
- * zero-sized array can only be placed at the tail of [nested]
- * structure definitions. lnet payload will be stored just after
- * the body of structure ksock_lnet_msg_t
- */
+ /* lnet hdr */
+ lnet_hdr_t ksnm_hdr;
} WIRE_ATTR ksock_lnet_msg_t;
typedef struct {
- __u32 ksm_type; /* type of socklnd message */
- __u32 ksm_csum; /* checksum if != 0 */
- __u64 ksm_zc_cookies[2]; /* Zero-Copy request/ACK cookie */
+ /* type of socklnd message */
+ __u32 ksm_type;
+ /* checksum if != 0 */
+ __u32 ksm_csum;
+ /* Zero-Copy request/ACK cookie */
+ __u64 ksm_zc_cookies[2];
union {
- ksock_lnet_msg_t lnetmsg; /* lnet message, it's empty if it's NOOP */
+ /* lnet message, it's empty if it's NOOP */
+ ksock_lnet_msg_t lnetmsg;
} WIRE_ATTR ksm_u;
} WIRE_ATTR ksock_msg_t;
static inline void
socklnd_init_msg(ksock_msg_t *msg, int type)
{
- msg->ksm_csum = 0;
- msg->ksm_type = type;
- msg->ksm_zc_cookies[0] = msg->ksm_zc_cookies[1] = 0;
+ msg->ksm_csum = 0;
+ msg->ksm_type = type;
+ msg->ksm_zc_cookies[0] = msg->ksm_zc_cookies[1] = 0;
}
-#define KSOCK_MSG_NOOP 0xc0 /* ksm_u empty */
-#define KSOCK_MSG_LNET 0xc1 /* lnet msg */
+#define KSOCK_MSG_NOOP 0xc0 /* ksm_u empty */
+#define KSOCK_MSG_LNET 0xc1 /* lnet msg */
/* We need to know this number to parse hello msg from ksocklnd in
* other LND (usocklnd, for example) */
-#define KSOCK_PROTO_V2 2
-#define KSOCK_PROTO_V3 3
+#define KSOCK_PROTO_V2 2
+#define KSOCK_PROTO_V3 3
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/types.h b/drivers/staging/lustre/include/linux/lnet/types.h
index 68d8139..53997ae 100644
--- a/drivers/staging/lustre/include/linux/lnet/types.h
+++ b/drivers/staging/lustre/include/linux/lnet/types.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,11 +23,11 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2014 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_TYPES_H__
@@ -48,7 +44,7 @@
/** Portal reserved for LNet's own use.
* \see lustre/include/lustre/lustre_idl.h for Lustre portal assignments.
*/
-#define LNET_RESERVED_PORTAL 0
+#define LNET_RESERVED_PORTAL 0
/**
* Address of an end-point in an LNet network.
@@ -74,8 +70,179 @@ typedef __u32 lnet_pid_t;
#define LNET_PID_RESERVED 0xf0000000 /* reserved bits in PID */
#define LNET_PID_USERFLAG 0x80000000 /* set in userspace peers */
+#define LNET_PID_LUSTRE 12345
+
+#define LNET_TIME_FOREVER (-1)
+
+/* how an LNET NID encodes net:address */
+/** extract the address part of an lnet_nid_t */
+
+static inline __u32 LNET_NIDADDR(lnet_nid_t nid)
+{
+ return nid & 0xffffffff;
+}
+
+static inline __u32 LNET_NIDNET(lnet_nid_t nid)
+{
+ return (nid >> 32) & 0xffffffff;
+}
+
+static inline lnet_nid_t LNET_MKNID(__u32 net, __u32 addr)
+{
+ return (((__u64)net) << 32) | addr;
+}
+
+static inline __u32 LNET_NETNUM(__u32 net)
+{
+ return net & 0xffff;
+}
+
+static inline __u32 LNET_NETTYP(__u32 net)
+{
+ return (net >> 16) & 0xffff;
+}
+
+static inline __u32 LNET_MKNET(__u32 type, __u32 num)
+{
+ return (type << 16) | num;
+}
+
+#define WIRE_ATTR __packed
+
+/* Packed version of lnet_process_id_t to transfer via network */
+typedef struct {
+ /* node id / process id */
+ lnet_nid_t nid;
+ lnet_pid_t pid;
+} WIRE_ATTR lnet_process_id_packed_t;
-#define LNET_TIME_FOREVER (-1)
+/* The wire handle's interface cookie only matches one network interface in
+ * one epoch (i.e. new cookie when the interface restarts or the node
+ * reboots). The object cookie only matches one object on that interface
+ * during that object's lifetime (i.e. no cookie re-use). */
+typedef struct {
+ __u64 wh_interface_cookie;
+ __u64 wh_object_cookie;
+} WIRE_ATTR lnet_handle_wire_t;
+
+typedef enum {
+ LNET_MSG_ACK = 0,
+ LNET_MSG_PUT,
+ LNET_MSG_GET,
+ LNET_MSG_REPLY,
+ LNET_MSG_HELLO,
+} lnet_msg_type_t;
+
+/* The variant fields of the portals message header are aligned on an 8
+ * byte boundary in the message header. Note that all types used in these
+ * wire structs MUST be fixed size and the smaller types are placed@the
+ * end. */
+typedef struct lnet_ack {
+ lnet_handle_wire_t dst_wmd;
+ __u64 match_bits;
+ __u32 mlength;
+} WIRE_ATTR lnet_ack_t;
+
+typedef struct lnet_put {
+ lnet_handle_wire_t ack_wmd;
+ __u64 match_bits;
+ __u64 hdr_data;
+ __u32 ptl_index;
+ __u32 offset;
+} WIRE_ATTR lnet_put_t;
+
+typedef struct lnet_get {
+ lnet_handle_wire_t return_wmd;
+ __u64 match_bits;
+ __u32 ptl_index;
+ __u32 src_offset;
+ __u32 sink_length;
+} WIRE_ATTR lnet_get_t;
+
+typedef struct lnet_reply {
+ lnet_handle_wire_t dst_wmd;
+} WIRE_ATTR lnet_reply_t;
+
+typedef struct lnet_hello {
+ __u64 incarnation;
+ __u32 type;
+} WIRE_ATTR lnet_hello_t;
+
+typedef struct {
+ lnet_nid_t dest_nid;
+ lnet_nid_t src_nid;
+ lnet_pid_t dest_pid;
+ lnet_pid_t src_pid;
+ __u32 type; /* lnet_msg_type_t */
+ __u32 payload_length; /* payload data to follow */
+ /*<------__u64 aligned------->*/
+ union {
+ lnet_ack_t ack;
+ lnet_put_t put;
+ lnet_get_t get;
+ lnet_reply_t reply;
+ lnet_hello_t hello;
+ } msg;
+} WIRE_ATTR lnet_hdr_t;
+
+/* A HELLO message contains a magic number and protocol version
+ * code in the header's dest_nid, the peer's NID in the src_nid, and
+ * LNET_MSG_HELLO in the type field. All other common fields are zero
+ * (including payload_size; i.e. no payload).
+ * This is for use by byte-stream LNDs (e.g. TCP/IP) to check the peer is
+ * running the same protocol and to find out its NID. These LNDs should
+ * exchange HELLO messages when a connection is first established. Individual
+ * LNDs can put whatever else they fancy in lnet_hdr_t::msg.
+ */
+typedef struct {
+ __u32 magic; /* LNET_PROTO_TCP_MAGIC */
+ __u16 version_major; /* increment on incompatible change */
+ __u16 version_minor; /* increment on compatible change */
+} WIRE_ATTR lnet_magicversion_t;
+
+/* PROTO MAGIC for LNDs */
+#define LNET_PROTO_IB_MAGIC 0x0be91b91
+#define LNET_PROTO_GNI_MAGIC 0xb00fbabe /* ask Kim */
+#define LNET_PROTO_TCP_MAGIC 0xeebc0ded
+#define LNET_PROTO_ACCEPTOR_MAGIC 0xacce7100
+#define LNET_PROTO_PING_MAGIC 0x70696E67 /* 'ping' */
+
+/* Placeholder for a future "unified" protocol across all LNDs */
+/* Current LNDs that receive a request with this magic will respond with a
+ * "stub" reply using their current protocol */
+#define LNET_PROTO_MAGIC 0x45726963 /* ! */
+
+#define LNET_PROTO_TCP_VERSION_MAJOR 1
+#define LNET_PROTO_TCP_VERSION_MINOR 0
+
+/* Acceptor connection request */
+typedef struct {
+ __u32 acr_magic; /* PTL_ACCEPTOR_PROTO_MAGIC */
+ __u32 acr_version; /* protocol version */
+ __u64 acr_nid; /* target NID */
+} WIRE_ATTR lnet_acceptor_connreq_t;
+
+#define LNET_PROTO_ACCEPTOR_VERSION 1
+
+typedef struct lnet_counters {
+ __u32 msgs_alloc;
+ __u32 msgs_max;
+ __u32 errors;
+ __u32 send_count;
+ __u32 recv_count;
+ __u32 route_count;
+ __u32 drop_count;
+ __u64 send_length;
+ __u64 recv_length;
+ __u64 route_length;
+ __u64 drop_length;
+} WIRE_ATTR lnet_counters_t;
+
+#define LNET_NI_STATUS_UP 0x15aac0de
+#define LNET_NI_STATUS_DOWN 0xdeadface
+#define LNET_NI_STATUS_INVALID 0x00000000
+
+#define LNET_MAX_INTERFACES 16
/**
* Objects maintained by the LNet are accessed through handles. Handle types
@@ -195,8 +362,8 @@ typedef struct {
* one must start on page boundary, and all but the last must end on
* page boundary.
*/
- void *start;
- unsigned int length;
+ void *start;
+ unsigned int length;
/**
* Specifies the maximum number of operations that can be performed
* on the memory descriptor. An operation is any action that could
@@ -207,7 +374,7 @@ typedef struct {
* there is no bound on the number of operations that may be applied
* to a MD.
*/
- int threshold;
+ int threshold;
/**
* Specifies the largest incoming request that the memory descriptor
* should respond to. When the unused portion of a MD (length -
@@ -215,7 +382,7 @@ typedef struct {
* does not respond to further operations. This value is only used
* if the LNET_MD_MAX_SIZE option is set.
*/
- int max_size;
+ int max_size;
/**
* Specifies the behavior of the memory descriptor. A bitwise OR
* of the following values can be used:
@@ -252,14 +419,14 @@ typedef struct {
* region (i.e. sum of all fragment lengths) must not be less than
* \a max_size.
*/
- unsigned int options;
+ unsigned int options;
/**
* A user-specified value that is associated with the memory
* descriptor. The value does not need to be a pointer, but must fit
* in the space used by a pointer. This value is recorded in events
* associated with operations on this MD.
*/
- void *user_ptr;
+ void *user_ptr;
/**
* A handle for the event queue used to log the operations performed on
* the memory region. If this argument is a NULL handle (i.e. nullified
@@ -276,44 +443,33 @@ typedef struct {
#define LNET_MTU (1 << LNET_MTU_BITS)
/** limit on the number of fragments in discontiguous MDs */
-#define LNET_MAX_IOV 256
-
-/* Max payload size */
-# define LNET_MAX_PAYLOAD CONFIG_LNET_MAX_PAYLOAD
-# if (LNET_MAX_PAYLOAD < LNET_MTU)
-# error "LNET_MAX_PAYLOAD too small - error in configure --with-max-payload-mb"
-# else
-# if (LNET_MAX_PAYLOAD > (PAGE_SIZE * LNET_MAX_IOV))
-/* PAGE_SIZE is a constant: check with cpp! */
-# error "LNET_MAX_PAYLOAD too large - error in configure --with-max-payload-mb"
-# endif
-# endif
+#define LNET_MAX_IOV 256
/**
* Options for the MD structure. See lnet_md_t::options.
*/
-#define LNET_MD_OP_PUT (1 << 0)
+#define LNET_MD_OP_PUT (1 << 0)
/** See lnet_md_t::options. */
-#define LNET_MD_OP_GET (1 << 1)
+#define LNET_MD_OP_GET (1 << 1)
/** See lnet_md_t::options. */
#define LNET_MD_MANAGE_REMOTE (1 << 2)
-/* unused (1 << 3) */
+/* unused (1 << 3) */
/** See lnet_md_t::options. */
-#define LNET_MD_TRUNCATE (1 << 4)
+#define LNET_MD_TRUNCATE (1 << 4)
/** See lnet_md_t::options. */
-#define LNET_MD_ACK_DISABLE (1 << 5)
+#define LNET_MD_ACK_DISABLE (1 << 5)
/** See lnet_md_t::options. */
#define LNET_MD_IOVEC (1 << 6)
/** See lnet_md_t::options. */
-#define LNET_MD_MAX_SIZE (1 << 7)
+#define LNET_MD_MAX_SIZE (1 << 7)
/** See lnet_md_t::options. */
-#define LNET_MD_KIOV (1 << 8)
+#define LNET_MD_KIOV (1 << 8)
/* For compatibility with Cray Portals */
-#define LNET_MD_PHYS 0
+#define LNET_MD_PHYS 0
/** Infinite threshold on MD operations. See lnet_md_t::threshold */
-#define LNET_MD_THRESH_INF (-1)
+#define LNET_MD_THRESH_INF (-1)
/* NB lustre portals uses struct iovec internally! */
typedef struct iovec lnet_md_iovec_t;
@@ -323,15 +479,15 @@ typedef struct iovec lnet_md_iovec_t;
*/
typedef struct {
/** Pointer to the page where the fragment resides */
- struct page *kiov_page;
+ struct page *kiov_page;
/** Length in bytes of the fragment */
- unsigned int kiov_len;
+ unsigned int kiov_len;
/**
* Starting offset of the fragment within the page. Note that the
* end of the fragment must not pass the end of the page; i.e.,
* kiov_len + kiov_offset <= PAGE_CACHE_SIZE.
*/
- unsigned int kiov_offset;
+ unsigned int kiov_offset;
} lnet_kiov_t;
/** @} lnet_md */
@@ -379,7 +535,7 @@ typedef enum {
LNET_EVENT_UNLINK,
} lnet_event_kind_t;
-#define LNET_SEQ_BASETYPE long
+#define LNET_SEQ_BASETYPE long
typedef unsigned LNET_SEQ_BASETYPE lnet_seq_t;
#define LNET_SEQ_GT(a, b) (((signed LNET_SEQ_BASETYPE)((a) - (b))) > 0)
@@ -388,23 +544,23 @@ typedef unsigned LNET_SEQ_BASETYPE lnet_seq_t;
*/
typedef struct {
/** The identifier (nid, pid) of the target. */
- lnet_process_id_t target;
+ lnet_process_id_t target;
/** The identifier (nid, pid) of the initiator. */
- lnet_process_id_t initiator;
+ lnet_process_id_t initiator;
/**
* The NID of the immediate sender. If the request has been forwarded
* by routers, this is the NID of the last hop; otherwise it's the
* same as the initiator.
*/
- lnet_nid_t sender;
+ lnet_nid_t sender;
/** Indicates the type of the event. */
- lnet_event_kind_t type;
+ lnet_event_kind_t type;
/** The portal table index specified in the request */
- unsigned int pt_index;
+ unsigned int pt_index;
/** A copy of the match bits specified in the request. */
- __u64 match_bits;
+ __u64 match_bits;
/** The length (in bytes) specified in the request. */
- unsigned int rlength;
+ unsigned int rlength;
/**
* The length (in bytes) of the data that was manipulated by the
* operation. For truncated operations, the manipulated length will be
@@ -412,47 +568,47 @@ typedef struct {
* see lnet_md_t). For all other operations, the manipulated length
* will be the length of the requested operation, i.e. rlength.
*/
- unsigned int mlength;
+ unsigned int mlength;
/**
* The handle to the MD associated with the event. The handle may be
* invalid if the MD has been unlinked.
*/
- lnet_handle_md_t md_handle;
+ lnet_handle_md_t md_handle;
/**
* A snapshot of the state of the MD immediately after the event has
* been processed. In particular, the threshold field in md will
* reflect the value of the threshold after the operation occurred.
*/
- lnet_md_t md;
+ lnet_md_t md;
/**
* 64 bits of out-of-band user data. Only valid for LNET_EVENT_PUT.
* \see LNetPut
*/
- __u64 hdr_data;
+ __u64 hdr_data;
/**
* Indicates the completion status of the operation. It's 0 for
* successful operations, otherwise it's an error code.
*/
- int status;
+ int status;
/**
* Indicates whether the MD has been unlinked. Note that:
* - An event with unlinked set is the last event on the MD.
* - This field is also set for an explicit LNET_EVENT_UNLINK event.
* \see LNetMDUnlink
*/
- int unlinked;
+ int unlinked;
/**
* The displacement (in bytes) into the memory region that the
* operation used. The offset can be determined by the operation for
* a remote managed MD or by the local MD.
* \see lnet_md_t::options
*/
- unsigned int offset;
+ unsigned int offset;
/**
* The sequence number for this event. Sequence numbers are unique
* to each event.
*/
- volatile lnet_seq_t sequence;
+ volatile lnet_seq_t sequence;
} lnet_event_t;
/**
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index be4de56..8a3341e 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -1177,7 +1177,7 @@ void kiblnd_map_rx_descs(kib_conn_t *conn)
CDEBUG(D_NET, "rx %d: %p %#llx(%#llx)\n",
i, rx->rx_msg, rx->rx_msgaddr,
- lnet_page2phys(pg) + pg_off);
+ (__u64)(page_to_phys(pg) + pg_off));
pg_off += IBLND_MSG_SIZE;
LASSERT(pg_off <= PAGE_SIZE);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
index cd664d0..32b39a4 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
@@ -62,16 +62,15 @@
#define DEBUG_SUBSYSTEM S_LND
-#include "../../../include/linux/libcfs/libcfs.h"
-#include "../../../include/linux/lnet/lnet.h"
-#include "../../../include/linux/lnet/lib-lnet.h"
-#include "../../../include/linux/lnet/lnet-sysctl.h"
-
#include <rdma/rdma_cm.h>
#include <rdma/ib_cm.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_fmr_pool.h>
+#include "../../../include/linux/libcfs/libcfs.h"
+#include "../../../include/linux/lnet/lnet.h"
+#include "../../../include/linux/lnet/lib-lnet.h"
+
#define IBLND_PEER_HASH_SIZE 101 /* # peer lists */
/* # scheduler loops before reschedule */
#define IBLND_RESCHED 100
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index f3b4923..8798e49 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -34,7 +34,6 @@
#include "../../../include/linux/lnet/lnet.h"
#include "../../../include/linux/lnet/lib-lnet.h"
#include "../../../include/linux/lnet/socklnd.h"
-#include "../../../include/linux/lnet/lnet-sysctl.h"
#define SOCKNAL_PEER_HASH_SIZE 101 /* # peer lists */
#define SOCKNAL_RESCHED 100 /* # scheduler loops before reschedule */
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index a93f59c..df0f7d8 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -148,8 +148,8 @@ ksocknal_lib_send_kiov(ksock_conn_t *conn, ksock_tx_t *tx)
rc = sk->sk_prot->sendpage(sk, page,
offset, fragsize, msgflg);
} else {
- rc = cfs_tcp_sendpage(sk, page, offset, fragsize,
- msgflg);
+ rc = tcp_sendpage(sk, page, offset, fragsize,
+ msgflg);
}
} else {
#if SOCKNAL_SINGLE_FRAG_TX || !SOCKNAL_RISK_KMAP_DEADLOCK
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index bd97099..c44d3e0 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -243,8 +243,6 @@ lnet_accept(struct socket *sock, __u32 magic)
if (magic == le32_to_cpu(LNET_PROTO_TCP_MAGIC))
str = "'old' socknal/tcpnal";
- else if (lnet_accept_magic(magic, LNET_PROTO_RA_MAGIC))
- str = "'old' ranal";
else
str = "unrecognised";
diff --git a/drivers/staging/lustre/lnet/lnet/api-ni.c b/drivers/staging/lustre/lnet/lnet/api-ni.c
index 4a14e51..f9ae9e5 100644
--- a/drivers/staging/lustre/lnet/lnet/api-ni.c
+++ b/drivers/staging/lustre/lnet/lnet/api-ni.c
@@ -364,56 +364,6 @@ lnet_counters_reset(void)
}
EXPORT_SYMBOL(lnet_counters_reset);
-#ifdef LNET_USE_LIB_FREELIST
-
-int
-lnet_freelist_init(lnet_freelist_t *fl, int n, int size)
-{
- char *space;
-
- LASSERT(n > 0);
-
- size += offsetof(lnet_freeobj_t, fo_contents);
-
- LIBCFS_ALLOC(space, n * size);
- if (space == NULL)
- return -ENOMEM;
-
- INIT_LIST_HEAD(&fl->fl_list);
- fl->fl_objs = space;
- fl->fl_nobjs = n;
- fl->fl_objsize = size;
-
- do {
- memset(space, 0, size);
- list_add((struct list_head *)space, &fl->fl_list);
- space += size;
- } while (--n != 0);
-
- return 0;
-}
-
-void
-lnet_freelist_fini(lnet_freelist_t *fl)
-{
- struct list_head *el;
- int count;
-
- if (fl->fl_nobjs == 0)
- return;
-
- count = 0;
- for (el = fl->fl_list.next; el != &fl->fl_list; el = el->next)
- count++;
-
- LASSERT(count == fl->fl_nobjs);
-
- LIBCFS_FREE(fl->fl_objs, fl->fl_nobjs * fl->fl_objsize);
- memset(fl, 0, sizeof(*fl));
-}
-
-#endif /* LNET_USE_LIB_FREELIST */
-
static __u64
lnet_create_interface_cookie(void)
{
@@ -470,9 +420,6 @@ lnet_res_container_cleanup(struct lnet_res_container *rec)
count, lnet_res_type2str(rec->rec_type));
}
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_fini(&rec->rec_freelist);
-#endif
if (rec->rec_lh_hash != NULL) {
LIBCFS_FREE(rec->rec_lh_hash,
LNET_LH_HASH_SIZE * sizeof(rec->rec_lh_hash[0]));
@@ -483,8 +430,7 @@ lnet_res_container_cleanup(struct lnet_res_container *rec)
}
static int
-lnet_res_container_setup(struct lnet_res_container *rec,
- int cpt, int type, int objnum, int objsz)
+lnet_res_container_setup(struct lnet_res_container *rec, int cpt, int type)
{
int rc = 0;
int i;
@@ -494,12 +440,6 @@ lnet_res_container_setup(struct lnet_res_container *rec,
rec->rec_type = type;
INIT_LIST_HEAD(&rec->rec_active);
-#ifdef LNET_USE_LIB_FREELIST
- memset(&rec->rec_freelist, 0, sizeof(rec->rec_freelist));
- rc = lnet_freelist_init(&rec->rec_freelist, objnum, objsz);
- if (rc != 0)
- goto out;
-#endif
rec->rec_lh_cookie = (cpt << LNET_COOKIE_TYPE_BITS) | type;
/* Arbitrary choice of hash table size */
@@ -535,7 +475,7 @@ lnet_res_containers_destroy(struct lnet_res_container **recs)
}
static struct lnet_res_container **
-lnet_res_containers_create(int type, int objnum, int objsz)
+lnet_res_containers_create(int type)
{
struct lnet_res_container **recs;
struct lnet_res_container *rec;
@@ -550,7 +490,7 @@ lnet_res_containers_create(int type, int objnum, int objsz)
}
cfs_percpt_for_each(rec, i, recs) {
- rc = lnet_res_container_setup(rec, i, type, objnum, objsz);
+ rc = lnet_res_container_setup(rec, i, type);
if (rc != 0) {
lnet_res_containers_destroy(recs);
return NULL;
@@ -597,7 +537,6 @@ lnet_res_lh_initialize(struct lnet_res_container *rec, lnet_libhandle_t *lh)
list_add(&lh->lh_hash_chain, &rec->rec_lh_hash[hash]);
}
-
int lnet_unprepare(void);
static int
@@ -643,13 +582,11 @@ lnet_prepare(lnet_pid_t requested_pid)
goto failed;
rc = lnet_res_container_setup(&the_lnet.ln_eq_container, 0,
- LNET_COOKIE_TYPE_EQ, LNET_FL_MAX_EQS,
- sizeof(lnet_eq_t));
+ LNET_COOKIE_TYPE_EQ);
if (rc != 0)
goto failed;
- recs = lnet_res_containers_create(LNET_COOKIE_TYPE_ME, LNET_FL_MAX_MES,
- sizeof(lnet_me_t));
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_ME);
if (recs == NULL) {
rc = -ENOMEM;
goto failed;
@@ -657,8 +594,7 @@ lnet_prepare(lnet_pid_t requested_pid)
the_lnet.ln_me_containers = recs;
- recs = lnet_res_containers_create(LNET_COOKIE_TYPE_MD, LNET_FL_MAX_MDS,
- sizeof(lnet_libmd_t));
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_MD);
if (recs == NULL) {
rc = -ENOMEM;
goto failed;
@@ -1171,15 +1107,14 @@ lnet_startup_lndnis(void)
/**
* Initialize LNet library.
*
- * Only userspace program needs to call this function - it's automatically
- * called in the kernel at module loading time. Caller has to call LNetFini()
- * after a call to LNetInit(), if and only if the latter returned 0. It must
- * be called exactly once.
+ * Automatically called at module loading time. Caller has to call
+ * lnet_fini() after a call to lnet_init(), if and only if the latter
+ * returned 0. It must be called exactly once.
*
* \return 0 on success, and -ve on failures.
*/
int
-LNetInit(void)
+lnet_init(void)
{
int rc;
@@ -1232,19 +1167,16 @@ LNetInit(void)
lnet_register_lnd(&the_lolnd);
return 0;
}
-EXPORT_SYMBOL(LNetInit);
+EXPORT_SYMBOL(lnet_init);
/**
* Finalize LNet library.
*
- * Only userspace program needs to call this function. It can be called
- * at most once.
- *
- * \pre LNetInit() called with success.
- * \pre All LNet users called LNetNIFini() for matching LNetNIInit() calls.
+ * \pre lnet_init() called with success.
+ * \pre All LNet users called lnet_fini() for matching lnet_init() calls.
*/
void
-LNetFini(void)
+lnet_fini(void)
{
LASSERT(the_lnet.ln_init);
LASSERT(the_lnet.ln_refcount == 0);
@@ -1256,7 +1188,7 @@ LNetFini(void)
the_lnet.ln_init = 0;
}
-EXPORT_SYMBOL(LNetFini);
+EXPORT_SYMBOL(lnet_fini);
/**
* Set LNet PID and start LNet interfaces, routing, and forwarding.
@@ -1290,8 +1222,6 @@ LNetNIInit(lnet_pid_t requested_pid)
goto out;
}
- lnet_get_tunables();
-
if (requested_pid == LNET_PID_ANY) {
/* Don't instantiate LNET just for me */
rc = -ENETDOWN;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-eq.c b/drivers/staging/lustre/lnet/lnet/lib-eq.c
index 5470148..fabca5d 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-eq.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-eq.c
@@ -191,7 +191,7 @@ LNetEQFree(lnet_handle_eq_t eqh)
lnet_res_lh_invalidate(&eq->eq_lh);
list_del(&eq->eq_list);
- lnet_eq_free_locked(eq);
+ lnet_eq_free(eq);
out:
lnet_eq_wait_unlock();
lnet_res_unlock(LNET_LOCK_EX);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-md.c b/drivers/staging/lustre/lnet/lnet/lib-md.c
index 89d660f..57ede1b 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-md.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-md.c
@@ -82,7 +82,7 @@ lnet_md_unlink(lnet_libmd_t *md)
LASSERT(!list_empty(&md->md_list));
list_del_init(&md->md_list);
- lnet_md_free_locked(md);
+ lnet_md_free(md);
}
static int
@@ -320,7 +320,7 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return 0;
failed:
- lnet_md_free_locked(md);
+ lnet_md_free(md);
lnet_res_unlock(cpt);
return rc;
@@ -381,7 +381,7 @@ LNetMDBind(lnet_md_t umd, lnet_unlink_t unlink, lnet_handle_md_t *handle)
return 0;
failed:
- lnet_md_free_locked(md);
+ lnet_md_free(md);
lnet_res_unlock(cpt);
return rc;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-me.c b/drivers/staging/lustre/lnet/lnet/lib-me.c
index a3f9292..823df95 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-me.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-me.c
@@ -172,7 +172,7 @@ LNetMEInsert(lnet_handle_me_t current_meh,
current_me = lnet_handle2me(¤t_meh);
if (current_me == NULL) {
- lnet_me_free_locked(new_me);
+ lnet_me_free(new_me);
lnet_res_unlock(cpt);
return -ENOENT;
@@ -183,7 +183,7 @@ LNetMEInsert(lnet_handle_me_t current_meh,
ptl = the_lnet.ln_portals[current_me->me_portal];
if (lnet_ptl_is_unique(ptl)) {
/* nosense to insertion on unique portal */
- lnet_me_free_locked(new_me);
+ lnet_me_free(new_me);
lnet_res_unlock(cpt);
return -EPERM;
}
@@ -276,23 +276,5 @@ lnet_me_unlink(lnet_me_t *me)
}
lnet_res_lh_invalidate(&me->me_lh);
- lnet_me_free_locked(me);
+ lnet_me_free(me);
}
-
-#if 0
-static void
-lib_me_dump(lnet_me_t *me)
-{
- CWARN("Match Entry %p (%#llx)\n", me,
- me->me_lh.lh_cookie);
-
- CWARN("\tMatch/Ignore\t= %016lx / %016lx\n",
- me->me_match_bits, me->me_ignore_bits);
-
- CWARN("\tMD\t= %p\n", me->md);
- CWARN("\tprev\t= %p\n",
- list_entry(me->me_list.prev, lnet_me_t, me_list));
- CWARN("\tnext\t= %p\n",
- list_entry(me->me_list.next, lnet_me_t, me_list));
-}
-#endif
diff --git a/drivers/staging/lustre/lnet/lnet/lib-msg.c b/drivers/staging/lustre/lnet/lnet/lib-msg.c
index a46ccbf..41a5900 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-msg.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-msg.c
@@ -427,7 +427,7 @@ lnet_complete_msg_locked(lnet_msg_t *msg, int cpt)
}
lnet_msg_decommit(msg, cpt, status);
- lnet_msg_free_locked(msg);
+ lnet_msg_free(msg);
return 0;
}
@@ -559,35 +559,17 @@ lnet_msg_container_cleanup(struct lnet_msg_container *container)
sizeof(*container->msc_finalizers));
container->msc_finalizers = NULL;
}
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_fini(&container->msc_freelist);
-#endif
container->msc_init = 0;
}
int
lnet_msg_container_setup(struct lnet_msg_container *container, int cpt)
{
- int rc;
-
container->msc_init = 1;
INIT_LIST_HEAD(&container->msc_active);
INIT_LIST_HEAD(&container->msc_finalizing);
-#ifdef LNET_USE_LIB_FREELIST
- memset(&container->msc_freelist, 0, sizeof(lnet_freelist_t));
-
- rc = lnet_freelist_init(&container->msc_freelist,
- LNET_FL_MAX_MSGS, sizeof(lnet_msg_t));
- if (rc != 0) {
- CERROR("Failed to init freelist for message container\n");
- lnet_msg_container_cleanup(container);
- return rc;
- }
-#else
- rc = 0;
-#endif
/* number of CPUs */
container->msc_nfinalizers = cfs_cpt_weight(lnet_cpt_table(), cpt);
@@ -601,7 +583,7 @@ lnet_msg_container_setup(struct lnet_msg_container *container, int cpt)
return -ENOMEM;
}
- return rc;
+ return 0;
}
void
diff --git a/drivers/staging/lustre/lnet/lnet/module.c b/drivers/staging/lustre/lnet/lnet/module.c
index 72b7fbc..bc74738 100644
--- a/drivers/staging/lustre/lnet/lnet/module.c
+++ b/drivers/staging/lustre/lnet/lnet/module.c
@@ -117,7 +117,7 @@ init_lnet(void)
mutex_init(&lnet_config_mutex);
- rc = LNetInit();
+ rc = lnet_init();
if (rc != 0) {
CERROR("LNetInit: error %d\n", rc);
return rc;
@@ -143,7 +143,7 @@ fini_lnet(void)
rc = libcfs_deregister_ioctl(&lnet_ioctl_handler);
LASSERT(rc == 0);
- LNetFini();
+ lnet_fini();
}
MODULE_AUTHOR("Peter J. Braam <braam@clusterfs.com>");
diff --git a/drivers/staging/lustre/lnet/lnet/router.c b/drivers/staging/lustre/lnet/lnet/router.c
index 8510bae..6acaaae 100644
--- a/drivers/staging/lustre/lnet/lnet/router.c
+++ b/drivers/staging/lustre/lnet/lnet/router.c
@@ -24,8 +24,6 @@
#define DEBUG_SUBSYSTEM S_LNET
#include "../../include/linux/lnet/lib-lnet.h"
-#if defined(LNET_ROUTER)
-
#define LNET_NRB_TINY_MIN 512 /* min value for each CPT */
#define LNET_NRB_TINY (LNET_NRB_TINY_MIN * 4)
#define LNET_NRB_SMALL_MIN 4096 /* min value for each CPT */
@@ -70,15 +68,6 @@ lnet_peer_buffer_credits(lnet_ni_t *ni)
/* forward ref's */
static int lnet_router_checker(void *);
-#else
-
-int
-lnet_peer_buffer_credits(lnet_ni_t *ni)
-{
- return 0;
-}
-
-#endif
static int check_routers_before_use;
module_param(check_routers_before_use, int, 0444);
@@ -1163,9 +1152,6 @@ lnet_prune_rc_data(int wait_unlink)
lnet_net_unlock(LNET_LOCK_EX);
}
-
-#if defined(LNET_ROUTER)
-
static int
lnet_router_checker(void *arg)
{
@@ -1573,134 +1559,3 @@ lnet_notify(lnet_ni_t *ni, lnet_nid_t nid, int alive, unsigned long when)
return 0;
}
EXPORT_SYMBOL(lnet_notify);
-
-void
-lnet_get_tunables(void)
-{
-}
-
-#else
-
-int
-lnet_notify(lnet_ni_t *ni, lnet_nid_t nid, int alive, unsigned long when)
-{
- return -EOPNOTSUPP;
-}
-
-void
-lnet_router_checker(void)
-{
- static time_t last;
- static int running;
-
- time_t now = get_seconds();
- int interval = now - last;
- int rc;
- __u64 version;
- lnet_peer_t *rtr;
-
- /* It's no use to call me again within a sec - all intervals and
- * timeouts are measured in seconds */
- if (last != 0 && interval < 2)
- return;
-
- if (last != 0 &&
- interval > max(live_router_check_interval,
- dead_router_check_interval))
- CNETERR("Checker(%d/%d) not called for %d seconds\n",
- live_router_check_interval, dead_router_check_interval,
- interval);
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_net_lock(0);
- LASSERT(!running); /* recursion check */
- running = 1;
- lnet_net_unlock(0);
-
- last = now;
-
- if (the_lnet.ln_rc_state == LNET_RC_STATE_STOPPING)
- lnet_prune_rc_data(0); /* unlink all rcd and nowait */
-
- /* consume all pending events */
- while (1) {
- int i;
- lnet_event_t ev;
-
- /* NB ln_rc_eqh must be the 1st in 'eventqs' otherwise the
- * recursion breaker in LNetEQPoll would fail */
- rc = LNetEQPoll(&the_lnet.ln_rc_eqh, 1, 0, &ev, &i);
- if (rc == 0) /* no event pending */
- break;
-
- /* NB a lost SENT prevents me from pinging a router again */
- if (rc == -EOVERFLOW) {
- CERROR("Dropped an event!!!\n");
- abort();
- }
-
- LASSERT(rc == 1);
-
- lnet_router_checker_event(&ev);
- }
-
- if (the_lnet.ln_rc_state == LNET_RC_STATE_STOPPING) {
- lnet_prune_rc_data(1); /* release rcd */
- the_lnet.ln_rc_state = LNET_RC_STATE_SHUTDOWN;
- running = 0;
- return;
- }
-
- LASSERT(the_lnet.ln_rc_state == LNET_RC_STATE_RUNNING);
-
- lnet_net_lock(0);
-
- version = the_lnet.ln_routers_version;
- list_for_each_entry(rtr, &the_lnet.ln_routers, lp_rtr_list) {
- lnet_ping_router_locked(rtr);
- LASSERT(version == the_lnet.ln_routers_version);
- }
-
- lnet_net_unlock(0);
-
- running = 0; /* lock only needed for the recursion check */
-}
-
-/* NB lnet_peers_start_down depends on me,
- * so must be called before any peer creation */
-void
-lnet_get_tunables(void)
-{
- char *s;
-
- s = getenv("LNET_ROUTER_PING_TIMEOUT");
- if (s != NULL)
- router_ping_timeout = atoi(s);
-
- s = getenv("LNET_LIVE_ROUTER_CHECK_INTERVAL");
- if (s != NULL)
- live_router_check_interval = atoi(s);
-
- s = getenv("LNET_DEAD_ROUTER_CHECK_INTERVAL");
- if (s != NULL)
- dead_router_check_interval = atoi(s);
-
- /* This replaces old lnd_notify mechanism */
- check_routers_before_use = 1;
- if (dead_router_check_interval <= 0)
- dead_router_check_interval = 30;
-}
-
-void
-lnet_rtrpools_free(void)
-{
-}
-
-int
-lnet_rtrpools_alloc(int im_a_arouter)
-{
- return 0;
-}
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index e2805bd..b32bc54 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -56,8 +56,8 @@
*/
#include "../../include/linux/libcfs/libcfs.h"
-// #include <obd.h>
-#include "../../include/linux/lnet/lnet.h"
+#include "../../include/linux/lnet/nidstr.h"
+#include "../../include/linux/lnet/api.h"
#include "lustre/lustre_idl.h"
#include "lustre_ha.h"
#include "lustre_sec.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons, James Simmons
From: James Simmons <uja.ornl@yahoo.com>
Currently the lnet headers used for user land applications
contain various kernel definations. This is due to the
fact libcfs contains kernel wrappers for user land which
will be going away. This patch sorted the header data
so all kernel containing structures are moved out of
headers that user land will use.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13792
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 4 -
.../lustre/include/linux/libcfs/libcfs_private.h | 51 --
.../lustre/include/linux/lnet/api-support.h | 44 --
drivers/staging/lustre/include/linux/lnet/api.h | 51 +-
.../staging/lustre/include/linux/lnet/lib-lnet.h | 249 +-------
.../staging/lustre/include/linux/lnet/lib-types.h | 649 +++++++++-----------
.../lustre/include/linux/lnet/linux/api-support.h | 42 --
.../lustre/include/linux/lnet/linux/lib-lnet.h | 71 ---
.../lustre/include/linux/lnet/linux/lib-types.h | 45 --
.../staging/lustre/include/linux/lnet/linux/lnet.h | 56 --
.../lustre/include/linux/lnet/lnet-sysctl.h | 49 --
drivers/staging/lustre/include/linux/lnet/lnet.h | 11 +-
.../staging/lustre/include/linux/lnet/lnetctl.h | 36 +-
drivers/staging/lustre/include/linux/lnet/lnetst.h | 665 ++++++++++++--------
drivers/staging/lustre/include/linux/lnet/nidstr.h | 73 +++
drivers/staging/lustre/include/linux/lnet/ptllnd.h | 93 ---
.../lustre/include/linux/lnet/ptllnd_wire.h | 119 ----
.../staging/lustre/include/linux/lnet/socklnd.h | 91 ++--
drivers/staging/lustre/include/linux/lnet/types.h | 266 +++++++--
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h | 9 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 1 -
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 4 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 2 -
drivers/staging/lustre/lnet/lnet/api-ni.c | 100 +---
drivers/staging/lustre/lnet/lnet/lib-eq.c | 2 +-
drivers/staging/lustre/lnet/lnet/lib-md.c | 6 +-
drivers/staging/lustre/lnet/lnet/lib-me.c | 24 +-
drivers/staging/lustre/lnet/lnet/lib-msg.c | 22 +-
drivers/staging/lustre/lnet/lnet/module.c | 4 +-
drivers/staging/lustre/lnet/lnet/router.c | 145 -----
drivers/staging/lustre/lustre/include/lustre_net.h | 4 +-
32 files changed, 1149 insertions(+), 1841 deletions(-)
delete mode 100644 drivers/staging/lustre/include/linux/lnet/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/api-support.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/linux/lnet.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
create mode 100644 drivers/staging/lustre/include/linux/lnet/nidstr.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd.h
delete mode 100644 drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index f54bab6..aae9487 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -46,10 +46,6 @@
#include "curproc.h"
-#ifndef offsetof
-# define offsetof(typ, memb) ((long)(long_ptr_t)((char *)&(((typ *)0)->memb)))
-#endif
-
#ifndef ARRAY_SIZE
#define ARRAY_SIZE(a) ((sizeof(a)) / (sizeof((a)[0])))
#endif
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
index fef8825..6174946 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs_private.h
@@ -410,36 +410,6 @@ int cfs_percpt_atomic_summary(atomic_t **refs);
*/
#define CLASSERT(cond) do {switch (42) {case (cond): case 0: break; } } while (0)
-/* support decl needed both by kernel and liblustre */
-int libcfs_isknown_lnd(int type);
-char *libcfs_lnd2modname(int type);
-char *libcfs_lnd2str(int type);
-int libcfs_str2lnd(const char *str);
-char *libcfs_net2str(__u32 net);
-char *libcfs_nid2str(lnet_nid_t nid);
-__u32 libcfs_str2net(const char *str);
-lnet_nid_t libcfs_str2nid(const char *str);
-int libcfs_str2anynid(lnet_nid_t *nid, const char *str);
-char *libcfs_id2str(lnet_process_id_t id);
-void cfs_free_nidlist(struct list_head *list);
-int cfs_parse_nidlist(char *str, int len, struct list_head *list);
-int cfs_match_nid(lnet_nid_t nid, struct list_head *list);
-
-/** \addtogroup lnet_addr
- * @{ */
-/* how an LNET NID encodes net:address */
-/** extract the address part of an lnet_nid_t */
-#define LNET_NIDADDR(nid) ((__u32)((nid) & 0xffffffff))
-/** extract the network part of an lnet_nid_t */
-#define LNET_NIDNET(nid) ((__u32)(((nid) >> 32)) & 0xffffffff)
-/** make an lnet_nid_t from a network part and an address part */
-#define LNET_MKNID(net, addr) ((((__u64)(net))<<32)|((__u64)(addr)))
-/* how net encodes type:number */
-#define LNET_NETNUM(net) ((net) & 0xffff)
-#define LNET_NETTYP(net) (((net) >> 16) & 0xffff)
-#define LNET_MKNET(typ, num) ((((__u32)(typ))<<16)|((__u32)(num)))
-/** @} lnet_addr */
-
/* max value for numeric network address */
#define MAX_NUMERIC_VALUE 0xffffffff
@@ -532,25 +502,4 @@ do { \
ptr += cfs_size_round(len + 1); \
} while (0)
-/**
- * Lustre Network Driver types.
- */
-enum {
- /* Only add to these values (i.e. don't ever change or redefine them):
- * network addresses depend on them... */
- QSWLND = 1,
- SOCKLND = 2,
- GMLND = 3, /* obsolete, keep it so that libcfs_nid2str works */
- PTLLND = 4,
- O2IBLND = 5,
- CIBLND = 6,
- OPENIBLND = 7,
- IIBLND = 8,
- LOLND = 9,
- RALND = 10,
- VIBLND = 11,
- MXLND = 12,
- GNILND = 13,
-};
-
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/api-support.h b/drivers/staging/lustre/include/linux/lnet/api-support.h
deleted file mode 100644
index 8f7fa28..0000000
--- a/drivers/staging/lustre/include/linux/lnet/api-support.h
+++ /dev/null
@@ -1,44 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_API_SUPPORT_H__
-#define __LNET_API_SUPPORT_H__
-
-#include "linux/api-support.h"
-
-#include "../libcfs/libcfs.h"
-#include "types.h"
-#include "lnet.h"
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/api.h b/drivers/staging/lustre/include/linux/lnet/api.h
index cd86517..76fb6fd 100644
--- a/drivers/staging/lustre/include/linux/lnet/api.h
+++ b/drivers/staging/lustre/include/linux/lnet/api.h
@@ -15,21 +15,19 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2011, 2015, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_API_H__
@@ -41,20 +39,15 @@
*
* LNet is an asynchronous message-passing API, which provides an unreliable
* connectionless service that can't guarantee any order. It supports OFA IB,
- * TCP/IP, and Cray Portals, and routes between heterogeneous networks.
- *
- * LNet can run both in OS kernel space and in userspace as a library.
+ * TCP/IP, and Cray Interconnects, and routes between heterogeneous networks.
* @{
*/
-#include "../lnet/types.h"
+#include "types.h"
/** \defgroup lnet_init_fini Initialization and cleanup
* The LNet must be properly initialized before any LNet calls can be made.
* @{ */
-int LNetInit(void);
-void LNetFini(void);
-
int LNetNIInit(lnet_pid_t requested_pid);
int LNetNIFini(void);
/** @} lnet_init_fini */
@@ -84,6 +77,7 @@ void LNetSnprintHandle(char *str, int str_len, lnet_handle_any_t handle);
/** @} lnet_addr */
+
/** \defgroup lnet_me Match entries
*
* A match entry (abbreviated as ME) describes a set of criteria to accept
@@ -98,16 +92,16 @@ void LNetSnprintHandle(char *str, int str_len, lnet_handle_any_t handle);
* @{ */
int LNetMEAttach(unsigned int portal,
lnet_process_id_t match_id_in,
- __u64 match_bits_in,
- __u64 ignore_bits_in,
+ __u64 match_bits_in,
+ __u64 ignore_bits_in,
lnet_unlink_t unlink_in,
lnet_ins_pos_t pos_in,
lnet_handle_me_t *handle_out);
int LNetMEInsert(lnet_handle_me_t current_in,
lnet_process_id_t match_id_in,
- __u64 match_bits_in,
- __u64 ignore_bits_in,
+ __u64 match_bits_in,
+ __u64 ignore_bits_in,
lnet_unlink_t unlink_in,
lnet_ins_pos_t position_in,
lnet_handle_me_t *handle_out);
@@ -128,11 +122,11 @@ int LNetMEUnlink(lnet_handle_me_t current_in);
* associated with a MD: LNetMDUnlink().
* @{ */
int LNetMDAttach(lnet_handle_me_t current_in,
- lnet_md_t md_in,
+ lnet_md_t md_in,
lnet_unlink_t unlink_in,
lnet_handle_md_t *handle_out);
-int LNetMDBind(lnet_md_t md_in,
+int LNetMDBind(lnet_md_t md_in,
lnet_unlink_t unlink_in,
lnet_handle_md_t *handle_out);
@@ -170,14 +164,15 @@ int LNetEQFree(lnet_handle_eq_t eventq_in);
int LNetEQGet(lnet_handle_eq_t eventq_in,
lnet_event_t *event_out);
+
int LNetEQWait(lnet_handle_eq_t eventq_in,
lnet_event_t *event_out);
int LNetEQPoll(lnet_handle_eq_t *eventqs_in,
- int neq_in,
- int timeout_ms,
+ int neq_in,
+ int timeout_ms,
lnet_event_t *event_out,
- int *which_eq_out);
+ int *which_eq_out);
/** @} lnet_eq */
/** \defgroup lnet_data Data movement operations
@@ -185,23 +180,24 @@ int LNetEQPoll(lnet_handle_eq_t *eventqs_in,
* The LNet API provides two data movement operations: LNetPut()
* and LNetGet().
* @{ */
-int LNetPut(lnet_nid_t self,
+int LNetPut(lnet_nid_t self,
lnet_handle_md_t md_in,
lnet_ack_req_t ack_req_in,
lnet_process_id_t target_in,
unsigned int portal_in,
- __u64 match_bits_in,
+ __u64 match_bits_in,
unsigned int offset_in,
- __u64 hdr_data_in);
+ __u64 hdr_data_in);
-int LNetGet(lnet_nid_t self,
+int LNetGet(lnet_nid_t self,
lnet_handle_md_t md_in,
lnet_process_id_t target_in,
unsigned int portal_in,
- __u64 match_bits_in,
+ __u64 match_bits_in,
unsigned int offset_in);
/** @} lnet_data */
+
/** \defgroup lnet_misc Miscellaneous operations.
* Miscellaneous operations.
* @{ */
@@ -209,7 +205,6 @@ int LNetGet(lnet_nid_t self,
int LNetSetLazyPortal(int portal);
int LNetClearLazyPortal(int portal);
int LNetCtl(unsigned int cmd, void *arg);
-int LNetSetAsync(lnet_process_id_t id, int nasync);
/** @} lnet_misc */
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index c00a2ca..5eacae4 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,56 +23,46 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2011, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lib-lnet.h
- *
- * Top level include for library side routines
*/
#ifndef __LNET_LIB_LNET_H__
#define __LNET_LIB_LNET_H__
-#include "linux/lib-lnet.h"
#include "../libcfs/libcfs.h"
-#include "types.h"
+#include "api.h"
#include "lnet.h"
#include "lib-types.h"
-extern lnet_t the_lnet; /* THE network */
-
-#if defined(LNET_USE_LIB_FREELIST)
-/* 1 CPT, simplify implementation... */
-# define LNET_CPT_MAX_BITS 0
-
-#else /* KERNEL and no freelist */
+extern lnet_t the_lnet; /* THE network */
-# if (BITS_PER_LONG == 32)
+#if (BITS_PER_LONG == 32)
/* 2 CPTs, allowing more CPTs might make us under memory pressure */
-# define LNET_CPT_MAX_BITS 1
+# define LNET_CPT_MAX_BITS 1
-# else /* 64-bit system */
+#else /* 64-bit system */
/*
* 256 CPTs for thousands of CPUs, allowing more CPTs might make us
* under risk of consuming all lh_cookie.
*/
-# define LNET_CPT_MAX_BITS 8
-# endif /* BITS_PER_LONG == 32 */
-#endif
+# define LNET_CPT_MAX_BITS 8
+#endif /* BITS_PER_LONG == 32 */
/* max allowed CPT number */
-#define LNET_CPT_MAX (1 << LNET_CPT_MAX_BITS)
+#define LNET_CPT_MAX (1 << LNET_CPT_MAX_BITS)
-#define LNET_CPT_NUMBER (the_lnet.ln_cpt_number)
-#define LNET_CPT_BITS (the_lnet.ln_cpt_bits)
-#define LNET_CPT_MASK ((1ULL << LNET_CPT_BITS) - 1)
+#define LNET_CPT_NUMBER (the_lnet.ln_cpt_number)
+#define LNET_CPT_BITS (the_lnet.ln_cpt_bits)
+#define LNET_CPT_MASK ((1ULL << LNET_CPT_BITS) - 1)
/** exclusive lock */
-#define LNET_LOCK_EX CFS_PERCPT_LOCK_EX
+#define LNET_LOCK_EX CFS_PERCPT_LOCK_EX
static inline int lnet_is_wire_handle_none(lnet_handle_wire_t *wh)
{
@@ -177,191 +163,9 @@ lnet_net_lock_current(void)
#define MAX_PORTALS 64
-/* these are only used by code with LNET_USE_LIB_FREELIST, but we still
- * exported them to !LNET_USE_LIB_FREELIST for easy implementation */
-#define LNET_FL_MAX_MES 2048
-#define LNET_FL_MAX_MDS 2048
-#define LNET_FL_MAX_EQS 512
-#define LNET_FL_MAX_MSGS 2048 /* Outstanding messages */
-
-#ifdef LNET_USE_LIB_FREELIST
-
-int lnet_freelist_init(lnet_freelist_t *fl, int n, int size);
-void lnet_freelist_fini(lnet_freelist_t *fl);
-
-static inline void *
-lnet_freelist_alloc(lnet_freelist_t *fl)
-{
- /* ALWAYS called with liblock held */
- lnet_freeobj_t *o;
-
- if (list_empty(&fl->fl_list))
- return NULL;
-
- o = list_entry(fl->fl_list.next, lnet_freeobj_t, fo_list);
- list_del(&o->fo_list);
- return (void *)&o->fo_contents;
-}
-
-static inline void
-lnet_freelist_free(lnet_freelist_t *fl, void *obj)
-{
- /* ALWAYS called with liblock held */
- lnet_freeobj_t *o = list_entry(obj, lnet_freeobj_t, fo_contents);
-
- list_add(&o->fo_list, &fl->fl_list);
-}
-
static inline lnet_eq_t *
lnet_eq_alloc(void)
{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = &the_lnet.ln_eq_container;
- lnet_eq_t *eq;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- eq = (lnet_eq_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- return eq;
-}
-
-static inline void
-lnet_eq_free_locked(lnet_eq_t *eq)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = &the_lnet.ln_eq_container;
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, eq);
-}
-
-static inline void
-lnet_eq_free(lnet_eq_t *eq)
-{
- lnet_res_lock(0);
- lnet_eq_free_locked(eq);
- lnet_res_unlock(0);
-}
-
-static inline lnet_libmd_t *
-lnet_md_alloc(lnet_md_t *umd)
-{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
- lnet_libmd_t *md;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- md = (lnet_libmd_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- if (md != NULL)
- INIT_LIST_HEAD(&md->md_list);
-
- return md;
-}
-
-static inline void
-lnet_md_free_locked(lnet_libmd_t *md)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, md);
-}
-
-static inline void
-lnet_md_free(lnet_libmd_t *md)
-{
- lnet_res_lock(0);
- lnet_md_free_locked(md);
- lnet_res_unlock(0);
-}
-
-static inline lnet_me_t *
-lnet_me_alloc(void)
-{
- /* NEVER called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
- lnet_me_t *me;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_res_lock(0);
- me = (lnet_me_t *)lnet_freelist_alloc(&rec->rec_freelist);
- lnet_res_unlock(0);
-
- return me;
-}
-
-static inline void
-lnet_me_free_locked(lnet_me_t *me)
-{
- /* ALWAYS called with resource lock held */
- struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- lnet_freelist_free(&rec->rec_freelist, me);
-}
-
-static inline void
-lnet_me_free(lnet_me_t *me)
-{
- lnet_res_lock(0);
- lnet_me_free_locked(me);
- lnet_res_unlock(0);
-}
-
-static inline lnet_msg_t *
-lnet_msg_alloc(void)
-{
- /* NEVER called with network lock held */
- struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
- lnet_msg_t *msg;
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_net_lock(0);
- msg = (lnet_msg_t *)lnet_freelist_alloc(&msc->msc_freelist);
- lnet_net_unlock(0);
-
- if (msg != NULL) {
- /* NULL pointers, clear flags etc */
- memset(msg, 0, sizeof(*msg));
- }
- return msg;
-}
-
-static inline void
-lnet_msg_free_locked(lnet_msg_t *msg)
-{
- /* ALWAYS called with network lock held */
- struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
-
- LASSERT(LNET_CPT_NUMBER == 1);
- LASSERT(!msg->msg_onactivelist);
- lnet_freelist_free(&msc->msc_freelist, msg);
-}
-
-static inline void
-lnet_msg_free(lnet_msg_t *msg)
-{
- lnet_net_lock(0);
- lnet_msg_free_locked(msg);
- lnet_net_unlock(0);
-}
-
-#else /* !LNET_USE_LIB_FREELIST */
-
-static inline lnet_eq_t *
-lnet_eq_alloc(void)
-{
- /* NEVER called with liblock held */
lnet_eq_t *eq;
LIBCFS_ALLOC(eq, sizeof(*eq));
@@ -371,14 +175,12 @@ lnet_eq_alloc(void)
static inline void
lnet_eq_free(lnet_eq_t *eq)
{
- /* ALWAYS called with resource lock held */
LIBCFS_FREE(eq, sizeof(*eq));
}
static inline lnet_libmd_t *
lnet_md_alloc(lnet_md_t *umd)
{
- /* NEVER called with liblock held */
lnet_libmd_t *md;
unsigned int size;
unsigned int niov;
@@ -407,7 +209,6 @@ lnet_md_alloc(lnet_md_t *umd)
static inline void
lnet_md_free(lnet_libmd_t *md)
{
- /* ALWAYS called with resource lock held */
unsigned int size;
if ((md->md_options & LNET_MD_KIOV) != 0)
@@ -421,7 +222,6 @@ lnet_md_free(lnet_libmd_t *md)
static inline lnet_me_t *
lnet_me_alloc(void)
{
- /* NEVER called with liblock held */
lnet_me_t *me;
LIBCFS_ALLOC(me, sizeof(*me));
@@ -431,14 +231,12 @@ lnet_me_alloc(void)
static inline void
lnet_me_free(lnet_me_t *me)
{
- /* ALWAYS called with resource lock held */
LIBCFS_FREE(me, sizeof(*me));
}
static inline lnet_msg_t *
lnet_msg_alloc(void)
{
- /* NEVER called with liblock held */
lnet_msg_t *msg;
LIBCFS_ALLOC(msg, sizeof(*msg));
@@ -450,18 +248,10 @@ lnet_msg_alloc(void)
static inline void
lnet_msg_free(lnet_msg_t *msg)
{
- /* ALWAYS called with network lock held */
LASSERT(!msg->msg_onactivelist);
LIBCFS_FREE(msg, sizeof(*msg));
}
-#define lnet_eq_free_locked(eq) lnet_eq_free(eq)
-#define lnet_md_free_locked(md) lnet_md_free(md)
-#define lnet_me_free_locked(me) lnet_me_free(me)
-#define lnet_msg_free_locked(msg) lnet_msg_free(msg)
-
-#endif /* LNET_USE_LIB_FREELIST */
-
lnet_libhandle_t *lnet_res_lh_lookup(struct lnet_res_container *rec,
__u64 cookie);
void lnet_res_lh_initialize(struct lnet_res_container *rec,
@@ -644,6 +434,9 @@ lnet_ni_t *lnet_nid2ni_locked(lnet_nid_t nid, int cpt);
lnet_ni_t *lnet_net2ni_locked(__u32 net, int cpt);
lnet_ni_t *lnet_net2ni(__u32 net);
+int lnet_init(void);
+void lnet_fini(void);
+
int lnet_notify(lnet_ni_t *ni, lnet_nid_t peer, int alive, unsigned long when);
void lnet_notify_locked(lnet_peer_t *lp, int notifylnd, int alive,
unsigned long when);
@@ -737,7 +530,9 @@ void lnet_recv(lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed,
unsigned int offset, unsigned int mlen, unsigned int rlen);
lnet_msg_t *lnet_create_reply_msg(lnet_ni_t *ni, lnet_msg_t *get_msg);
void lnet_set_reply_msg_len(lnet_ni_t *ni, lnet_msg_t *msg, unsigned int len);
+
void lnet_finalize(lnet_ni_t *ni, lnet_msg_t *msg, int rc);
+
void lnet_drop_delayed_msg_list(struct list_head *head, char *reason);
void lnet_recv_delayed_msg_list(struct list_head *head);
@@ -887,12 +682,12 @@ void lnet_peer_tables_destroy(void);
int lnet_peer_tables_create(void);
void lnet_debug_peer(lnet_nid_t nid);
-static inline void lnet_peer_set_alive(lnet_peer_t *lp)
+static inline void
+lnet_peer_set_alive(lnet_peer_t *lp)
{
lp->lp_last_alive = lp->lp_last_query = get_seconds();
if (!lp->lp_alive)
lnet_notify_locked(lp, 0, 1, lp->lp_last_alive);
}
-
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-types.h b/drivers/staging/lustre/include/linux/lnet/lib-types.h
index 5053766..87acf14 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-types.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-types.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,155 +23,42 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lib-types.h
- *
- * Types used by the library side routines that do not need to be
- * exposed to the user application
*/
#ifndef __LNET_LIB_TYPES_H__
#define __LNET_LIB_TYPES_H__
-#include "linux/lib-types.h"
-
-#include "../libcfs/libcfs.h"
-#include <linux/list.h>
-#include "types.h"
-
-#define WIRE_ATTR __attribute__((packed))
-
-/* Packed version of lnet_process_id_t to transfer via network */
-typedef struct {
- lnet_nid_t nid;
- lnet_pid_t pid; /* node id / process id */
-} WIRE_ATTR lnet_process_id_packed_t;
-
-/* The wire handle's interface cookie only matches one network interface in
- * one epoch (i.e. new cookie when the interface restarts or the node
- * reboots). The object cookie only matches one object on that interface
- * during that object's lifetime (i.e. no cookie re-use). */
-typedef struct {
- __u64 wh_interface_cookie;
- __u64 wh_object_cookie;
-} WIRE_ATTR lnet_handle_wire_t;
-
-typedef enum {
- LNET_MSG_ACK = 0,
- LNET_MSG_PUT,
- LNET_MSG_GET,
- LNET_MSG_REPLY,
- LNET_MSG_HELLO,
-} lnet_msg_type_t;
-
-/* The variant fields of the portals message header are aligned on an 8
- * byte boundary in the message header. Note that all types used in these
- * wire structs MUST be fixed size and the smaller types are placed at the
- * end. */
-typedef struct lnet_ack {
- lnet_handle_wire_t dst_wmd;
- __u64 match_bits;
- __u32 mlength;
-} WIRE_ATTR lnet_ack_t;
-
-typedef struct lnet_put {
- lnet_handle_wire_t ack_wmd;
- __u64 match_bits;
- __u64 hdr_data;
- __u32 ptl_index;
- __u32 offset;
-} WIRE_ATTR lnet_put_t;
-
-typedef struct lnet_get {
- lnet_handle_wire_t return_wmd;
- __u64 match_bits;
- __u32 ptl_index;
- __u32 src_offset;
- __u32 sink_length;
-} WIRE_ATTR lnet_get_t;
-
-typedef struct lnet_reply {
- lnet_handle_wire_t dst_wmd;
-} WIRE_ATTR lnet_reply_t;
-
-typedef struct lnet_hello {
- __u64 incarnation;
- __u32 type;
-} WIRE_ATTR lnet_hello_t;
+#include <linux/kthread.h>
+#include <linux/uio.h>
+#include <linux/types.h>
+#include <net/sock.h>
-typedef struct {
- lnet_nid_t dest_nid;
- lnet_nid_t src_nid;
- lnet_pid_t dest_pid;
- lnet_pid_t src_pid;
- __u32 type; /* lnet_msg_type_t */
- __u32 payload_length; /* payload data to follow */
- /*<------__u64 aligned------->*/
- union {
- lnet_ack_t ack;
- lnet_put_t put;
- lnet_get_t get;
- lnet_reply_t reply;
- lnet_hello_t hello;
- } msg;
-} WIRE_ATTR lnet_hdr_t;
-
-/* A HELLO message contains a magic number and protocol version
- * code in the header's dest_nid, the peer's NID in the src_nid, and
- * LNET_MSG_HELLO in the type field. All other common fields are zero
- * (including payload_size; i.e. no payload).
- * This is for use by byte-stream LNDs (e.g. TCP/IP) to check the peer is
- * running the same protocol and to find out its NID. These LNDs should
- * exchange HELLO messages when a connection is first established. Individual
- * LNDs can put whatever else they fancy in lnet_hdr_t::msg.
- */
-typedef struct {
- __u32 magic; /* LNET_PROTO_TCP_MAGIC */
- __u16 version_major; /* increment on incompatible change */
- __u16 version_minor; /* increment on compatible change */
-} WIRE_ATTR lnet_magicversion_t;
-
-/* PROTO MAGIC for LNDs */
-#define LNET_PROTO_IB_MAGIC 0x0be91b91
-#define LNET_PROTO_RA_MAGIC 0x0be91b92
-#define LNET_PROTO_QSW_MAGIC 0x0be91b93
-#define LNET_PROTO_GNI_MAGIC 0xb00fbabe /* ask Kim */
-#define LNET_PROTO_TCP_MAGIC 0xeebc0ded
-#define LNET_PROTO_PTL_MAGIC 0x50746C4E /* 'PtlN' unique magic */
-#define LNET_PROTO_MX_MAGIC 0x4d583130 /* 'MX10'! */
-#define LNET_PROTO_ACCEPTOR_MAGIC 0xacce7100
-#define LNET_PROTO_PING_MAGIC 0x70696E67 /* 'ping' */
-
-/* Placeholder for a future "unified" protocol across all LNDs */
-/* Current LNDs that receive a request with this magic will respond with a
- * "stub" reply using their current protocol */
-#define LNET_PROTO_MAGIC 0x45726963 /* ! */
-
-#define LNET_PROTO_TCP_VERSION_MAJOR 1
-#define LNET_PROTO_TCP_VERSION_MINOR 0
-
-/* Acceptor connection request */
-typedef struct {
- __u32 acr_magic; /* PTL_ACCEPTOR_PROTO_MAGIC */
- __u32 acr_version; /* protocol version */
- __u64 acr_nid; /* target NID */
-} WIRE_ATTR lnet_acceptor_connreq_t;
+#include "lnetctl.h"
-#define LNET_PROTO_ACCEPTOR_VERSION 1
+/* Max payload size */
+# define LNET_MAX_PAYLOAD CONFIG_LNET_MAX_PAYLOAD
+# if (LNET_MAX_PAYLOAD < LNET_MTU)
+# error "LNET_MAX_PAYLOAD too small - error in configure --with-max-payload-mb"
+# elif (LNET_MAX_PAYLOAD > (PAGE_SIZE * LNET_MAX_IOV))
+# error "LNET_MAX_PAYLOAD too large - error in configure --with-max-payload-mb"
+# endif
/* forward refs */
struct lnet_libmd;
typedef struct lnet_msg {
- struct list_head msg_activelist;
- struct list_head msg_list; /* Q for credits/MD */
+ struct list_head msg_activelist;
+ /* Q for credits/MD */
+ struct list_head msg_list;
- lnet_process_id_t msg_target;
+ lnet_process_id_t msg_target;
/* where is it from, it's only for building event */
lnet_nid_t msg_from;
__u32 msg_type;
@@ -195,110 +78,115 @@ typedef struct lnet_msg {
/* ready for pending on RX delay list */
unsigned int msg_rx_ready_delay:1;
- unsigned int msg_vmflush:1; /* VM trying to free memory */
- unsigned int msg_target_is_router:1; /* sending to a router */
- unsigned int msg_routing:1; /* being forwarded */
- unsigned int msg_ack:1; /* ack on finalize (PUT) */
- unsigned int msg_sending:1; /* outgoing message */
- unsigned int msg_receiving:1; /* being received */
- unsigned int msg_txcredit:1; /* taken an NI send credit */
- unsigned int msg_peertxcredit:1; /* taken a peer send credit */
- unsigned int msg_rtrcredit:1; /* taken a global router credit */
- unsigned int msg_peerrtrcredit:1; /* taken a peer router credit */
- unsigned int msg_onactivelist:1; /* on the activelist */
-
- struct lnet_peer *msg_txpeer; /* peer I'm sending to */
- struct lnet_peer *msg_rxpeer; /* peer I received from */
-
- void *msg_private;
- struct lnet_libmd *msg_md;
-
- unsigned int msg_len;
- unsigned int msg_wanted;
- unsigned int msg_offset;
- unsigned int msg_niov;
- struct kvec *msg_iov;
- lnet_kiov_t *msg_kiov;
-
- lnet_event_t msg_ev;
- lnet_hdr_t msg_hdr;
+ /* VM trying to free memory */
+ unsigned int msg_vmflush:1;
+ /* sending to a router */
+ unsigned int msg_target_is_router:1;
+ /* being forwarded */
+ unsigned int msg_routing:1;
+ /* ack on finalize (PUT) */
+ unsigned int msg_ack:1;
+ /* outgoing message */
+ unsigned int msg_sending:1;
+ /* being received */
+ unsigned int msg_receiving:1;
+ /* taken an NI send credit */
+ unsigned int msg_txcredit:1;
+ /* taken a peer send credit */
+ unsigned int msg_peertxcredit:1;
+ /* taken a global router credit */
+ unsigned int msg_rtrcredit:1;
+ /* taken a peer router credit */
+ unsigned int msg_peerrtrcredit:1;
+ /* on the activelist */
+ unsigned int msg_onactivelist:1;
+
+ /* peer I'm sending to */
+ struct lnet_peer *msg_txpeer;
+ /* peer I received from */
+ struct lnet_peer *msg_rxpeer;
+
+ void *msg_private;
+ struct lnet_libmd *msg_md;
+
+ unsigned int msg_len;
+ unsigned int msg_wanted;
+ unsigned int msg_offset;
+ unsigned int msg_niov;
+ struct kvec *msg_iov;
+ lnet_kiov_t *msg_kiov;
+
+ lnet_event_t msg_ev;
+ lnet_hdr_t msg_hdr;
} lnet_msg_t;
typedef struct lnet_libhandle {
- struct list_head lh_hash_chain;
- __u64 lh_cookie;
+ struct list_head lh_hash_chain;
+ __u64 lh_cookie;
} lnet_libhandle_t;
#define lh_entry(ptr, type, member) \
((type *)((char *)(ptr)-(char *)(&((type *)0)->member)))
typedef struct lnet_eq {
- struct list_head eq_list;
+ struct list_head eq_list;
lnet_libhandle_t eq_lh;
lnet_seq_t eq_enq_seq;
lnet_seq_t eq_deq_seq;
unsigned int eq_size;
lnet_eq_handler_t eq_callback;
lnet_event_t *eq_events;
- int **eq_refs; /* percpt refcount for EQ */
+ /* percpt refcount for EQ */
+ int **eq_refs;
} lnet_eq_t;
typedef struct lnet_me {
- struct list_head me_list;
- lnet_libhandle_t me_lh;
- lnet_process_id_t me_match_id;
- unsigned int me_portal;
- unsigned int me_pos; /* hash offset in mt_hash */
- __u64 me_match_bits;
- __u64 me_ignore_bits;
- lnet_unlink_t me_unlink;
- struct lnet_libmd *me_md;
+ struct list_head me_list;
+ lnet_libhandle_t me_lh;
+ lnet_process_id_t me_match_id;
+ unsigned int me_portal;
+ /* hash offset in mt_hash */
+ unsigned int me_pos;
+ __u64 me_match_bits;
+ __u64 me_ignore_bits;
+ lnet_unlink_t me_unlink;
+ struct lnet_libmd *me_md;
} lnet_me_t;
typedef struct lnet_libmd {
- struct list_head md_list;
- lnet_libhandle_t md_lh;
- lnet_me_t *md_me;
- char *md_start;
- unsigned int md_offset;
- unsigned int md_length;
- unsigned int md_max_size;
- int md_threshold;
- int md_refcount;
- unsigned int md_options;
- unsigned int md_flags;
- void *md_user_ptr;
- lnet_eq_t *md_eq;
- unsigned int md_niov; /* # frags */
+ struct list_head md_list;
+ lnet_libhandle_t md_lh;
+ lnet_me_t *md_me;
+ char *md_start;
+ unsigned int md_offset;
+ unsigned int md_length;
+ unsigned int md_max_size;
+ int md_threshold;
+ int md_refcount;
+ unsigned int md_options;
+ unsigned int md_flags;
+ void *md_user_ptr;
+ lnet_eq_t *md_eq;
+ /* # frags */
+ unsigned int md_niov;
union {
- struct kvec iov[LNET_MAX_IOV];
- lnet_kiov_t kiov[LNET_MAX_IOV];
+ struct kvec iov[LNET_MAX_IOV];
+ lnet_kiov_t kiov[LNET_MAX_IOV];
} md_iov;
} lnet_libmd_t;
-#define LNET_MD_FLAG_ZOMBIE (1 << 0)
-#define LNET_MD_FLAG_AUTO_UNLINK (1 << 1)
-#define LNET_MD_FLAG_ABORTED (1 << 2)
+#define LNET_MD_FLAG_ZOMBIE (1 << 0)
+#define LNET_MD_FLAG_AUTO_UNLINK (1 << 1)
+#define LNET_MD_FLAG_ABORTED (1 << 2)
-#ifdef LNET_USE_LIB_FREELIST
+/* info about peers we are trying to fail */
typedef struct {
- void *fl_objs; /* single contiguous array of objects */
- int fl_nobjs; /* the number of them */
- int fl_objsize; /* the size (including overhead) of each of them */
- struct list_head fl_list; /* where they are enqueued */
-} lnet_freelist_t;
-
-typedef struct {
- struct list_head fo_list; /* enqueue on fl_list */
- void *fo_contents; /* aligned contents */
-} lnet_freeobj_t;
-#endif
-
-typedef struct {
- /* info about peers we are trying to fail */
- struct list_head tp_list; /* ln_test_peers */
- lnet_nid_t tp_nid; /* matching nid */
- unsigned int tp_threshold; /* # failures to simulate */
+ /* ln_test_peers */
+ struct list_head tp_list;
+ /* matching nid */
+ lnet_nid_t tp_nid;
+ /* # failures to simulate */
+ unsigned int tp_threshold;
} lnet_test_peer_t;
#define LNET_COOKIE_TYPE_MD 1
@@ -307,15 +195,17 @@ typedef struct {
#define LNET_COOKIE_TYPE_BITS 2
#define LNET_COOKIE_MASK ((1ULL << LNET_COOKIE_TYPE_BITS) - 1ULL)
-struct lnet_ni; /* forward ref */
+struct lnet_ni; /* forward ref */
+/* fields managed by portals */
typedef struct lnet_lnd {
- /* fields managed by portals */
- struct list_head lnd_list; /* stash in the LND table */
- int lnd_refcount; /* # active instances */
+ /* stash in the LND table */
+ struct list_head lnd_list;
+ /* # active instances */
+ int lnd_refcount;
/* fields initialised by the LND */
- unsigned int lnd_type;
+ __u32 lnd_type;
int (*lnd_startup)(struct lnet_ni *ni);
void (*lnd_shutdown)(struct lnet_ni *ni);
@@ -366,47 +256,57 @@ typedef struct lnet_lnd {
/* accept a new connection */
int (*lnd_accept)(struct lnet_ni *ni, struct socket *sock);
-
} lnd_t;
-#define LNET_NI_STATUS_UP 0x15aac0de
-#define LNET_NI_STATUS_DOWN 0xdeadface
-#define LNET_NI_STATUS_INVALID 0x00000000
typedef struct {
- lnet_nid_t ns_nid;
- __u32 ns_status;
- __u32 ns_unused;
+ lnet_nid_t ns_nid;
+ __u32 ns_status;
+ __u32 ns_unused;
} WIRE_ATTR lnet_ni_status_t;
struct lnet_tx_queue {
- int tq_credits; /* # tx credits free */
- int tq_credits_min; /* lowest it's been */
- int tq_credits_max; /* total # tx credits */
- struct list_head tq_delayed; /* delayed TXs */
+ /* # tx credits free */
+ int tq_credits;
+ /* lowest it's been */
+ int tq_credits_min;
+ /* total # tx credits */
+ int tq_credits_max;
+ /* delayed TXs */
+ struct list_head tq_delayed;
};
-#define LNET_MAX_INTERFACES 16
-
typedef struct lnet_ni {
spinlock_t ni_lock;
- struct list_head ni_list; /* chain on ln_nis */
- struct list_head ni_cptlist; /* chain on ln_nis_cpt */
- int ni_maxtxcredits; /* # tx credits */
+ /* chain on ln_nis */
+ struct list_head ni_list;
+ /* chain on ln_nis_cpt */
+ struct list_head ni_cptlist;
+ /* # tx credits */
+ int ni_maxtxcredits;
/* # per-peer send credits */
int ni_peertxcredits;
/* # per-peer router buffer credits */
int ni_peerrtrcredits;
/* seconds to consider peer dead */
int ni_peertimeout;
- int ni_ncpts; /* number of CPTs */
- __u32 *ni_cpts; /* bond NI on some CPTs */
- lnet_nid_t ni_nid; /* interface's NID */
- void *ni_data; /* instance-specific data */
- lnd_t *ni_lnd; /* procedural interface */
- struct lnet_tx_queue **ni_tx_queues; /* percpt TX queues */
- int **ni_refs; /* percpt reference count */
- long ni_last_alive; /* when I was last alive */
- lnet_ni_status_t *ni_status; /* my health status */
+ /* number of CPTs */
+ int ni_ncpts;
+ /* bond NI on some CPTs */
+ __u32 *ni_cpts;
+ /* interface's NID */
+ lnet_nid_t ni_nid;
+ /* instance-specific data */
+ void *ni_data;
+ /* procedural interface */
+ lnd_t *ni_lnd;
+ /* percpt TX queues */
+ struct lnet_tx_queue **ni_tx_queues;
+ /* percpt reference count */
+ int **ni_refs;
+ /* when I was last alive */
+ long ni_last_alive;
+ /* my health status */
+ lnet_ni_status_t *ni_status;
/* equivalent interfaces to use */
char *ni_interfaces[LNET_MAX_INTERFACES];
} lnet_ni_t;
@@ -435,43 +335,72 @@ typedef struct {
#define LNET_PINGINFO_SIZE offsetof(lnet_ping_info_t, pi_ni[LNET_MAX_RTR_NIS])
typedef struct {
/* chain on the_lnet.ln_zombie_rcd or ln_deathrow_rcd */
- struct list_head rcd_list;
- lnet_handle_md_t rcd_mdh; /* ping buffer MD */
- struct lnet_peer *rcd_gateway; /* reference to gateway */
- lnet_ping_info_t *rcd_pinginfo; /* ping buffer */
+ struct list_head rcd_list;
+ /* ping buffer MD */
+ lnet_handle_md_t rcd_mdh;
+ /* reference to gateway */
+ struct lnet_peer *rcd_gateway;
+ /* ping buffer */
+ lnet_ping_info_t *rcd_pinginfo;
} lnet_rc_data_t;
typedef struct lnet_peer {
- struct list_head lp_hashlist; /* chain on peer hash */
- struct list_head lp_txq; /* messages blocking for tx credits */
- struct list_head lp_rtrq; /* messages blocking for router credits */
- struct list_head lp_rtr_list; /* chain on router list */
- int lp_txcredits; /* # tx credits available */
- int lp_mintxcredits; /* low water mark */
- int lp_rtrcredits; /* # router credits */
- int lp_minrtrcredits; /* low water mark */
- unsigned int lp_alive:1; /* alive/dead? */
- unsigned int lp_notify:1; /* notification outstanding? */
- unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
- unsigned int lp_notifying:1; /* some thread is handling notification */
- unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
- int lp_alive_count; /* # times router went dead<->alive */
- long lp_txqnob; /* bytes queued for sending */
- unsigned long lp_timestamp; /* time of last aliveness news */
- unsigned long lp_ping_timestamp; /* time of last ping attempt */
- unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
- unsigned long lp_last_alive; /* when I was last alive */
- unsigned long lp_last_query; /* when lp_ni was queried last time */
- lnet_ni_t *lp_ni; /* interface peer is on */
- lnet_nid_t lp_nid; /* peer's NID */
- int lp_refcount; /* # refs */
- int lp_cpt; /* CPT this peer attached on */
+ /* chain on peer hash */
+ struct list_head lp_hashlist;
+ /* messages blocking for tx credits */
+ struct list_head lp_txq;
+ /* messages blocking for router credits */
+ struct list_head lp_rtrq;
+ /* chain on router list */
+ struct list_head lp_rtr_list;
+ /* # tx credits available */
+ int lp_txcredits;
+ /* low water mark */
+ int lp_mintxcredits;
+ /* # router credits */
+ int lp_rtrcredits;
+ /* low water mark */
+ int lp_minrtrcredits;
+ /* alive/dead? */
+ unsigned int lp_alive:1;
+ /* notification outstanding? */
+ unsigned int lp_notify:1;
+ /* outstanding notification for LND? */
+ unsigned int lp_notifylnd:1;
+ /* some thread is handling notification */
+ unsigned int lp_notifying:1;
+ /* SEND event outstanding from ping */
+ unsigned int lp_ping_notsent;
+ /* # times router went dead<->alive */
+ int lp_alive_count;
+ /* bytes queued for sending */
+ long lp_txqnob;
+ /* time of last aliveness news */
+ unsigned long lp_timestamp;
+ /* time of last ping attempt */
+ unsigned long lp_ping_timestamp;
+ /* != 0 if ping reply expected */
+ unsigned long lp_ping_deadline;
+ /* when I was last alive */
+ unsigned long lp_last_alive;
+ /* when lp_ni was queried last time */
+ unsigned long lp_last_query;
+ /* interface peer is on */
+ lnet_ni_t *lp_ni;
+ /* peer's NID */
+ lnet_nid_t lp_nid;
+ /* # refs */
+ int lp_refcount;
+ /* CPT this peer attached on */
+ int lp_cpt;
/* # refs from lnet_route_t::lr_gateway */
int lp_rtr_refcount;
/* returned RC ping features */
unsigned int lp_ping_feats;
- struct list_head lp_routes; /* routers on this peer */
- lnet_rc_data_t *lp_rcd; /* router checker state */
+ /* routers on this peer */
+ struct list_head lp_routes;
+ /* router checker state */
+ lnet_rc_data_t *lp_rcd;
} lnet_peer_t;
/* peer hash size */
@@ -480,10 +409,14 @@ typedef struct lnet_peer {
/* peer hash table */
struct lnet_peer_table {
- int pt_version; /* /proc validity stamp */
- int pt_number; /* # peers extant */
- struct list_head pt_deathrow; /* zombie peers */
- struct list_head *pt_hash; /* NID->peer hash */
+ /* /proc validity stamp */
+ int pt_version;
+ /* # peers extant */
+ int pt_number;
+ /* zombie peers */
+ struct list_head pt_deathrow;
+ /* NID->peer hash */
+ struct list_head *pt_hash;
};
/* peer aliveness is enabled only on routers for peers in a network where the
@@ -492,14 +425,22 @@ struct lnet_peer_table {
(lp)->lp_ni->ni_peertimeout > 0)
typedef struct {
- struct list_head lr_list; /* chain on net */
- struct list_head lr_gwlist; /* chain on gateway */
- lnet_peer_t *lr_gateway; /* router node */
- __u32 lr_net; /* remote network number */
- int lr_seq; /* sequence for round-robin */
- unsigned int lr_downis; /* number of down NIs */
- unsigned int lr_hops; /* how far I am */
- unsigned int lr_priority; /* route priority */
+ /* chain on net */
+ struct list_head lr_list;
+ /* chain on gateway */
+ struct list_head lr_gwlist;
+ /* router node */
+ lnet_peer_t *lr_gateway;
+ /* remote network number */
+ __u32 lr_net;
+ /* sequence for round-robin */
+ int lr_seq;
+ /* number of down NIs */
+ unsigned int lr_downis;
+ /* how far I am */
+ unsigned int lr_hops;
+ /* route priority */
+ unsigned int lr_priority;
} lnet_route_t;
#define LNET_REMOTE_NETS_HASH_DEFAULT (1U << 7)
@@ -507,43 +448,41 @@ typedef struct {
#define LNET_REMOTE_NETS_HASH_SIZE (1 << the_lnet.ln_remote_nets_hbits)
typedef struct {
- struct list_head lrn_list; /* chain on ln_remote_nets_hash */
- struct list_head lrn_routes; /* routes to me */
- __u32 lrn_net; /* my net number */
+ /* chain on ln_remote_nets_hash */
+ struct list_head lrn_list;
+ /* routes to me */
+ struct list_head lrn_routes;
+ /* my net number */
+ __u32 lrn_net;
} lnet_remotenet_t;
typedef struct {
- struct list_head rbp_bufs; /* my free buffer pool */
- struct list_head rbp_msgs; /* messages blocking for a buffer */
- int rbp_npages; /* # pages in each buffer */
- int rbp_nbuffers; /* # buffers */
- int rbp_credits; /* # free buffers / blocked messages */
- int rbp_mincredits; /* low water mark */
+ /* my free buffer pool */
+ struct list_head rbp_bufs;
+ /* messages blocking for a buffer */
+ struct list_head rbp_msgs;
+ /* # pages in each buffer */
+ int rbp_npages;
+ /* # buffers */
+ int rbp_nbuffers;
+ /* # free buffers / blocked messages */
+ int rbp_credits;
+ /* low water mark */
+ int rbp_mincredits;
} lnet_rtrbufpool_t;
typedef struct {
- struct list_head rb_list; /* chain on rbp_bufs */
- lnet_rtrbufpool_t *rb_pool; /* owning pool */
- lnet_kiov_t rb_kiov[0]; /* the buffer space */
+ /* chain on rbp_bufs */
+ struct list_head rb_list;
+ /* owning pool */
+ lnet_rtrbufpool_t *rb_pool;
+ /* the buffer space */
+ lnet_kiov_t rb_kiov[0];
} lnet_rtrbuf_t;
-typedef struct {
- __u32 msgs_alloc;
- __u32 msgs_max;
- __u32 errors;
- __u32 send_count;
- __u32 recv_count;
- __u32 route_count;
- __u32 drop_count;
- __u64 send_length;
- __u64 recv_length;
- __u64 route_length;
- __u64 drop_length;
-} WIRE_ATTR lnet_counters_t;
-
-#define LNET_PEER_HASHSIZE 503 /* prime! */
-
-#define LNET_NRBPOOLS 3 /* # different router buffer pools */
+#define LNET_PEER_HASHSIZE 503 /* prime! */
+
+#define LNET_NRBPOOLS 3 /* # different router buffer pools */
enum {
/* Didn't match anything */
@@ -559,9 +498,9 @@ enum {
};
/* Options for lnet_portal_t::ptl_options */
-#define LNET_PTL_LAZY (1 << 0)
-#define LNET_PTL_MATCH_UNIQUE (1 << 1) /* unique match, for RDMA */
-#define LNET_PTL_MATCH_WILDCARD (1 << 2) /* wildcard match, request portal */
+#define LNET_PTL_LAZY (1 << 0)
+#define LNET_PTL_MATCH_UNIQUE (1 << 1) /* unique match, for RDMA */
+#define LNET_PTL_MATCH_WILDCARD (1 << 2) /* wildcard match, request portal */
/* parameter for matching operations (GET, PUT) */
struct lnet_match_info {
@@ -591,13 +530,15 @@ struct lnet_match_info {
struct lnet_match_table {
/* reserved for upcoming patches, CPU partition ID */
unsigned int mt_cpt;
- unsigned int mt_portal; /* portal index */
+ /* portal index */
+ unsigned int mt_portal;
/* match table is set as "enabled" if there's non-exhausted MD
* attached on mt_mhash, it's only valid for wildcard portal */
unsigned int mt_enabled;
/* bitmap to flag whether MEs on mt_hash are exhausted or not */
__u64 mt_exhausted[LNET_MT_EXHAUSTED_BMAP];
- struct list_head *mt_mhash; /* matching hash */
+ /* matching hash */
+ struct list_head *mt_mhash;
};
/* these are only useful for wildcard portal */
@@ -612,21 +553,22 @@ struct lnet_match_table {
typedef struct lnet_portal {
spinlock_t ptl_lock;
- unsigned int ptl_index; /* portal ID, reserved */
+ /* portal ID, reserved */
+ unsigned int ptl_index;
/* flags on this portal: lazy, unique... */
unsigned int ptl_options;
/* list of messages which are stealing buffer */
- struct list_head ptl_msg_stealing;
+ struct list_head ptl_msg_stealing;
/* messages blocking for MD */
- struct list_head ptl_msg_delayed;
+ struct list_head ptl_msg_delayed;
/* Match table for each CPT */
struct lnet_match_table **ptl_mtables;
/* spread rotor of incoming "PUT" */
unsigned int ptl_rotor;
/* # active entries for this portal */
- int ptl_mt_nmaps;
+ int ptl_mt_nmaps;
/* array of active entries' cpu-partition-id */
- int ptl_mt_maps[0];
+ int ptl_mt_maps[0];
} lnet_portal_t;
#define LNET_LH_HASH_BITS 12
@@ -635,28 +577,28 @@ typedef struct lnet_portal {
/* resource container (ME, MD, EQ) */
struct lnet_res_container {
- unsigned int rec_type; /* container type */
- __u64 rec_lh_cookie; /* cookie generator */
- struct list_head rec_active; /* active resource list */
- struct list_head *rec_lh_hash; /* handle hash */
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_t rec_freelist; /* freelist for resources */
-#endif
+ /* container type */
+ unsigned int rec_type;
+ /* cookie generator */
+ __u64 rec_lh_cookie;
+ /* active resource list */
+ struct list_head rec_active;
+ /* handle hash */
+ struct list_head *rec_lh_hash;
};
/* message container */
struct lnet_msg_container {
- int msc_init; /* initialized or not */
+ /* initialized or not */
+ int msc_init;
/* max # threads finalizing */
int msc_nfinalizers;
/* msgs waiting to complete finalizing */
- struct list_head msc_finalizing;
- struct list_head msc_active; /* active message list */
+ struct list_head msc_finalizing;
+ /* active message list */
+ struct list_head msc_active;
/* threads doing finalization */
void **msc_finalizers;
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_t msc_freelist; /* freelist for messages */
-#endif
};
/* Router Checker states */
@@ -684,7 +626,7 @@ typedef struct {
/* Event Queue container */
struct lnet_res_container ln_eq_container;
- wait_queue_head_t ln_eq_waitq;
+ wait_queue_head_t ln_eq_waitq;
spinlock_t ln_eq_wait_lock;
unsigned int ln_remote_nets_hbits;
@@ -695,23 +637,25 @@ typedef struct {
lnet_counters_t **ln_counters;
struct lnet_peer_table **ln_peer_tables;
/* failure simulation */
- struct list_head ln_test_peers;
+ struct list_head ln_test_peers;
- struct list_head ln_nis; /* LND instances */
+ /* LND instances */
+ struct list_head ln_nis;
/* NIs bond on specific CPT(s) */
- struct list_head ln_nis_cpt;
+ struct list_head ln_nis_cpt;
/* dying LND instances */
- struct list_head ln_nis_zombie;
- lnet_ni_t *ln_loni; /* the loopback NI */
+ struct list_head ln_nis_zombie;
+ /* the loopback NI */
+ lnet_ni_t *ln_loni;
/* NI to wait for events in */
lnet_ni_t *ln_eq_waitni;
/* remote networks with routes to them */
- struct list_head *ln_remote_nets_hash;
+ struct list_head *ln_remote_nets_hash;
/* validity stamp */
__u64 ln_remote_nets_version;
/* list of all known routers */
- struct list_head ln_routers;
+ struct list_head ln_routers;
/* validity stamp */
__u64 ln_routers_version;
/* percpt router buffer pools */
@@ -726,15 +670,16 @@ typedef struct {
/* router checker's event queue */
lnet_handle_eq_t ln_rc_eqh;
/* rcd still pending on net */
- struct list_head ln_rcd_deathrow;
+ struct list_head ln_rcd_deathrow;
/* rcd ready for free */
- struct list_head ln_rcd_zombie;
+ struct list_head ln_rcd_zombie;
/* serialise startup/shutdown */
struct semaphore ln_rc_signal;
struct mutex ln_api_mutex;
struct mutex ln_lnd_mutex;
- int ln_init; /* LNetInit() called? */
+ /* LNetInit() called? */
+ int ln_init;
/* Have I called LNetNIInit myself? */
int ln_niinit_self;
/* LNetNIInit/LNetNIFini counter */
@@ -742,12 +687,14 @@ typedef struct {
/* shutdown in progress */
int ln_shutdown;
- int ln_routing; /* am I a router? */
- lnet_pid_t ln_pid; /* requested pid */
+ /* am I a router? */
+ int ln_routing;
+ /* requested pid */
+ lnet_pid_t ln_pid;
/* uniquely identifies this ni in this epoch */
__u64 ln_interface_cookie;
/* registered LNDs */
- struct list_head ln_lnds;
+ struct list_head ln_lnds;
/* space for network names */
char *ln_network_tokens;
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/api-support.h b/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
deleted file mode 100644
index e237ad6..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
+++ /dev/null
@@ -1,42 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LINUX_API_SUPPORT_H__
-#define __LINUX_API_SUPPORT_H__
-
-#ifndef __LNET_API_SUPPORT_H__
-#error Do not #include this file directly. #include <lnet /api-support.h> instead
-#endif
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
deleted file mode 100644
index 0f8f04d..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
+++ /dev/null
@@ -1,71 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LIB_LNET_H__
-#define __LNET_LINUX_LIB_LNET_H__
-
-#ifndef __LNET_LIB_LNET_H__
-#error Do not #include this file directly. #include <linux/lnet/lib-lnet.h> instead
-#endif
-
-# include <asm/page.h>
-# include <linux/string.h>
-# include <asm/io.h>
-#include "../../libcfs/libcfs.h"
-
-static inline __u64
-lnet_page2phys(struct page *p)
-{
- /* compiler optimizer will elide unused branches */
-
- switch (sizeof(typeof(page_to_phys(p)))) {
- case 4:
- /* page_to_phys returns a 32 bit physical address. This must
- * be a 32 bit machine with <= 4G memory and we must ensure we
- * don't sign extend when converting to 64 bits. */
- return (unsigned long)page_to_phys(p);
-
- case 8:
- /* page_to_phys returns a 64 bit physical address :) */
- return page_to_phys(p);
-
- default:
- LBUG();
- return 0;
- }
-}
-
-#define LNET_ROUTER
-
-#endif /* __LNET_LINUX_LIB_LNET_H__ */
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
deleted file mode 100644
index 669e8c0..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
+++ /dev/null
@@ -1,45 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LIB_TYPES_H__
-#define __LNET_LINUX_LIB_TYPES_H__
-
-#ifndef __LNET_LIB_TYPES_H__
-#error Do not #include this file directly. #include <linux/lnet/lib-types.h> instead
-#endif
-
-# include <linux/uio.h>
-# include <linux/types.h>
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
deleted file mode 100644
index 1e888f1..0000000
--- a/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
+++ /dev/null
@@ -1,56 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- *
- * Copyright (c) 2011, Intel Corporation.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_LINUX_LNET_H__
-#define __LNET_LINUX_LNET_H__
-
-#ifndef __LNET_H__
-#error Do not #include this file directly. #include <linux/lnet/lnet.h> instead
-#endif
-
-/*
- * lnet.h
- *
- * User application interface file
- */
-
-#include <linux/uio.h>
-#include <linux/types.h>
-
-#define cfs_tcp_sendpage(sk, page, offset, size, flags) \
- tcp_sendpage(sk, page, offset, size, flags)
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h b/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
deleted file mode 100644
index 2dee1b9..0000000
--- a/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
+++ /dev/null
@@ -1,49 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- */
-
-#ifndef __LNET_SYSCTL_H__
-#define __LNET_SYSCTL_H__
-
-#if defined(CONFIG_SYSCTL)
-
-#define CTL_KRANAL 201
-#define CTL_O2IBLND 205
-#define CTL_PTLLND 206
-#define CTL_QSWNAL 207
-#define CTL_SOCKLND 208
-#define CTL_GNILND 210
-
-#endif
-
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet.h b/drivers/staging/lustre/include/linux/lnet/lnet.h
index 75c0ab9..a31b29c 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnet.h
@@ -26,10 +26,12 @@
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_H__
@@ -40,12 +42,7 @@
*
* User application interface file
*/
-#include "linux/lnet.h"
-
#include "types.h"
-#include "api.h"
-
-#define LNET_NIDSTR_COUNT 1024 /* # of nidstrings */
-#define LNET_NIDSTR_SIZE 32 /* size of each one (see below for usage) */
+#include "nidstr.h"
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetctl.h b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
index 98181d3..2b82a06 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnetctl.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
@@ -14,27 +14,29 @@
* along with Portals; if not, write to the Free Software
* Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*
- * header for libptlctl.a
+ * header for lnet ioctl
*/
-#ifndef _PTLCTL_H_
-#define _PTLCTL_H_
+/*
+ * Copyright (c) 2014, Intel Corporation.
+ */
+#ifndef _LNETCTL_H_
+#define _LNETCTL_H_
-#include "../libcfs/libcfs.h"
#include "types.h"
-#define LNET_DEV_ID 0
-#define LNET_DEV_PATH "/dev/lnet"
-#define LNET_DEV_MAJOR 10
-#define LNET_DEV_MINOR 240
-#define OBD_DEV_ID 1
-#define OBD_DEV_NAME "obd"
-#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME
-#define OBD_DEV_MAJOR 10
-#define OBD_DEV_MINOR 241
-#define SMFS_DEV_ID 2
-#define SMFS_DEV_PATH "/dev/snapdev"
-#define SMFS_DEV_MAJOR 10
-#define SMFS_DEV_MINOR 242
+#define LNET_DEV_ID 0
+#define LNET_DEV_PATH "/dev/lnet"
+#define LNET_DEV_MAJOR 10
+#define LNET_DEV_MINOR 240
+#define OBD_DEV_ID 1
+#define OBD_DEV_NAME "obd"
+#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME
+#define OBD_DEV_MAJOR 10
+#define OBD_DEV_MINOR 241
+#define SMFS_DEV_ID 2
+#define SMFS_DEV_PATH "/dev/snapdev"
+#define SMFS_DEV_MAJOR 10
+#define SMFS_DEV_MINOR 242
int ptl_initialize(int argc, char **argv);
int jt_ptl_network(int argc, char **argv);
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetst.h b/drivers/staging/lustre/include/linux/lnet/lnetst.h
index 885f708..462eae8 100644
--- a/drivers/staging/lustre/include/linux/lnet/lnetst.h
+++ b/drivers/staging/lustre/include/linux/lnet/lnetst.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,123 +23,182 @@
* Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, Intel Corporation.
+ * Copyright (c) 2012, 2014, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/lnetst.h
*
- * Author: Liang Zhen <liangzhen@clusterfs.com>
+ * Author: Liang Zhen <liang.zhen@intel.com>
*/
#ifndef __LNET_ST_H__
#define __LNET_ST_H__
-#include "../libcfs/libcfs.h"
-#include "lnet.h"
-#include "lib-types.h"
-
#define LST_FEAT_NONE (0)
#define LST_FEAT_BULK_LEN (1 << 0) /* enable variable page size */
#define LST_FEATS_EMPTY (LST_FEAT_NONE)
#define LST_FEATS_MASK (LST_FEAT_NONE | LST_FEAT_BULK_LEN)
-#define LST_NAME_SIZE 32 /* max name buffer length */
-
-#define LSTIO_DEBUG 0xC00 /* debug */
-#define LSTIO_SESSION_NEW 0xC01 /* create session */
-#define LSTIO_SESSION_END 0xC02 /* end session */
-#define LSTIO_SESSION_INFO 0xC03 /* query session */
-#define LSTIO_GROUP_ADD 0xC10 /* add group */
-#define LSTIO_GROUP_LIST 0xC11 /* list all groups in session */
-#define LSTIO_GROUP_INFO 0xC12 /* query default information of specified group */
-#define LSTIO_GROUP_DEL 0xC13 /* delete group */
-#define LSTIO_NODES_ADD 0xC14 /* add nodes to specified group */
-#define LSTIO_GROUP_UPDATE 0xC15 /* update group */
-#define LSTIO_BATCH_ADD 0xC20 /* add batch */
-#define LSTIO_BATCH_START 0xC21 /* start batch */
-#define LSTIO_BATCH_STOP 0xC22 /* stop batch */
-#define LSTIO_BATCH_DEL 0xC23 /* delete batch */
-#define LSTIO_BATCH_LIST 0xC24 /* show all batches in the session */
-#define LSTIO_BATCH_INFO 0xC25 /* show defail of specified batch */
-#define LSTIO_TEST_ADD 0xC26 /* add test (to batch) */
-#define LSTIO_BATCH_QUERY 0xC27 /* query batch status */
-#define LSTIO_STAT_QUERY 0xC30 /* get stats */
-
-typedef struct {
- lnet_nid_t ses_nid; /* nid of console node */
- __u64 ses_stamp; /* time stamp */
-} lst_sid_t; /*** session id */
+#define LST_NAME_SIZE 32 /* max name buffer length */
+
+/* debug */
+#define LSTIO_DEBUG 0xC00
+/* create session */
+#define LSTIO_SESSION_NEW 0xC01
+/* end session */
+#define LSTIO_SESSION_END 0xC02
+/* query session */
+#define LSTIO_SESSION_INFO 0xC03
+/* add group */
+#define LSTIO_GROUP_ADD 0xC10
+/* list all groups in session */
+#define LSTIO_GROUP_LIST 0xC11
+/* query default information of specified group */
+#define LSTIO_GROUP_INFO 0xC12
+/* delete group */
+#define LSTIO_GROUP_DEL 0xC13
+/* add nodes to specified group */
+#define LSTIO_NODES_ADD 0xC14
+/* update group */
+#define LSTIO_GROUP_UPDATE 0xC15
+/* add batch */
+#define LSTIO_BATCH_ADD 0xC20
+/* start batch */
+#define LSTIO_BATCH_START 0xC21
+/* stop batch */
+#define LSTIO_BATCH_STOP 0xC22
+/* delete batch */
+#define LSTIO_BATCH_DEL 0xC23
+/* show all batches in the session */
+#define LSTIO_BATCH_LIST 0xC24
+/* show details of specified batch */
+#define LSTIO_BATCH_INFO 0xC25
+/* add test (to batch) */
+#define LSTIO_TEST_ADD 0xC26
+/* query batch status */
+#define LSTIO_BATCH_QUERY 0xC27
+/* get stats */
+#define LSTIO_STAT_QUERY 0xC30
+
+/*** session id */
+typedef struct {
+ /* nid of console node */
+ lnet_nid_t ses_nid;
+ /* time stamp */
+ __u64 ses_stamp;
+} lst_sid_t;
extern lst_sid_t LST_INVALID_SID;
-typedef struct {
- __u64 bat_id; /* unique id in session */
-} lst_bid_t; /*** batch id (group of tests) */
-
-/* Status of test node */
-#define LST_NODE_ACTIVE 0x1 /* node in this session */
-#define LST_NODE_BUSY 0x2 /* node is taken by other session */
-#define LST_NODE_DOWN 0x4 /* node is down */
-#define LST_NODE_UNKNOWN 0x8 /* node not in session */
-
-typedef struct {
- lnet_process_id_t nde_id; /* id of node */
- int nde_state; /* state of node */
-} lstcon_node_ent_t; /*** node entry, for list_group command */
-
-typedef struct {
- int nle_nnode; /* # of nodes */
- int nle_nactive; /* # of active nodes */
- int nle_nbusy; /* # of busy nodes */
- int nle_ndown; /* # of down nodes */
- int nle_nunknown; /* # of unknown nodes */
-} lstcon_ndlist_ent_t; /*** node_list entry, for list_batch command */
-
-typedef struct {
- int tse_type; /* test type */
- int tse_loop; /* loop count */
- int tse_concur; /* concurrency of test */
-} lstcon_test_ent_t; /*** test summary entry, for list_batch command */
-
-typedef struct {
- int bae_state; /* batch status */
- int bae_timeout; /* batch timeout */
- int bae_ntest; /* # of tests in the batch */
-} lstcon_batch_ent_t; /*** batch summary entry, for list_batch command */
-
-typedef struct {
- lstcon_ndlist_ent_t tbe_cli_nle; /* client (group) node_list entry */
- lstcon_ndlist_ent_t tbe_srv_nle; /* server (group) node_list entry */
+/*** batch id (group of tests) */
+typedef struct {
+ /* unique id in session */
+ __u64 bat_id;
+} lst_bid_t;
+
+/*** Status of test node */
+
+/* node in this session */
+#define LST_NODE_ACTIVE 0x1
+/* node is taken by other session */
+#define LST_NODE_BUSY 0x2
+/* node is down */
+#define LST_NODE_DOWN 0x4
+/* node not in session */
+#define LST_NODE_UNKNOWN 0x8
+
+/*** node entry, for list_group command */
+typedef struct {
+ /* id of node */
+ lnet_process_id_t nde_id;
+ /* state of node */
+ int nde_state;
+} lstcon_node_ent_t;
+
+/*** node_list entry, for list_batch command */
+typedef struct {
+ /* # of nodes */
+ int nle_nnode;
+ /* # of active nodes */
+ int nle_nactive;
+ /* # of busy nodes */
+ int nle_nbusy;
+ /* # of down nodes */
+ int nle_ndown;
+ /* # of unknown nodes */
+ int nle_nunknown;
+} lstcon_ndlist_ent_t;
+
+/*** test summary entry, for list_batch command */
+typedef struct {
+ /* test type */
+ int tse_type;
+ /* loop count */
+ int tse_loop;
+ /* concurrency of test */
+ int tse_concur;
+} lstcon_test_ent_t;
+
+/*** batch summary entry, for list_batch command */
+typedef struct {
+ /* batch status */
+ int bae_state;
+ /* batch status */
+ int bae_timeout;
+ /* # of tests in the batch */
+ int bae_ntest;
+} lstcon_batch_ent_t;
+
+/*** test/batch verbose information entry, for list_batch command */
+typedef struct {
+ /* client (group) node_list entry */
+ lstcon_ndlist_ent_t tbe_cli_nle;
+ /* server (group) node_list entry */
+ lstcon_ndlist_ent_t tbe_srv_nle;
union {
- lstcon_test_ent_t tbe_test; /* test entry */
- lstcon_batch_ent_t tbe_batch; /* batch entry */
+ /* test entry */
+ lstcon_test_ent_t tbe_test;
+ /* batch entry */
+ lstcon_batch_ent_t tbe_batch;
} u;
-} lstcon_test_batch_ent_t; /*** test/batch verbose information entry,
- *** for list_batch command */
-
-typedef struct {
- struct list_head rpe_link; /* link chain */
- lnet_process_id_t rpe_peer; /* peer's id */
- struct timeval rpe_stamp; /* time stamp of RPC */
- int rpe_state; /* peer's state */
- int rpe_rpc_errno; /* RPC errno */
-
- lst_sid_t rpe_sid; /* peer's session id */
- int rpe_fwk_errno; /* framework errno */
- int rpe_priv[4]; /* private data */
- char rpe_payload[0]; /* private reply payload */
+} lstcon_test_batch_ent_t;
+
+typedef struct {
+ /* link chain */
+ struct list_head rpe_link;
+ /* peer's id */
+ lnet_process_id_t rpe_peer;
+ /* time stamp of RPC */
+ struct timeval rpe_stamp;
+ /* peer's state */
+ int rpe_state;
+ /* RPC errno */
+ int rpe_rpc_errno;
+ /* peer's session id */
+ lst_sid_t rpe_sid;
+ /* framework errno */
+ int rpe_fwk_errno;
+ /* private data */
+ int rpe_priv[4];
+ /* private reply payload */
+ char rpe_payload[0];
} lstcon_rpc_ent_t;
typedef struct {
- int trs_rpc_stat[4]; /* RPCs stat (0: total, 1: failed, 2: finished, 4: reserved */
- int trs_rpc_errno; /* RPC errno */
- int trs_fwk_stat[8]; /* framework stat */
- int trs_fwk_errno; /* errno of the first remote error */
- void *trs_fwk_private; /* private framework stat */
+ /* RPCs stat (0: total, 1: failed, 2: finished, 4: reserved */
+ int trs_rpc_stat[4];
+ /* RPC errno */
+ int trs_rpc_errno;
+ /* framework stat */
+ int trs_fwk_stat[8];
+ /* errno of the first remote error */
+ int trs_fwk_errno;
+ /* private framework stat */
+ void *trs_fwk_private;
} lstcon_trans_stat_t;
static inline int
@@ -238,237 +293,351 @@ lstcon_statqry_stat_failure(lstcon_trans_stat_t *stat, int inc)
/* create a session */
typedef struct {
- int lstio_ses_key; /* IN: local key */
- int lstio_ses_timeout; /* IN: session timeout */
- int lstio_ses_force; /* IN: force create ? */
+ /* IN: local key */
+ int lstio_ses_key;
+ /* IN: session timeout */
+ int lstio_ses_timeout;
+ /* IN: force create ? */
+ int lstio_ses_force;
/** IN: session features */
- unsigned lstio_ses_feats;
- lst_sid_t *lstio_ses_idp; /* OUT: session id */
- int lstio_ses_nmlen; /* IN: name length */
- char *lstio_ses_namep; /* IN: session name */
+ unsigned lstio_ses_feats;
+ /* OUT: session id */
+ lst_sid_t *lstio_ses_idp;
+ /* IN: name length */
+ int lstio_ses_nmlen;
+ /* IN: session name */
+ char *lstio_ses_namep;
} lstio_session_new_args_t;
/* query current session */
typedef struct {
- lst_sid_t *lstio_ses_idp; /* OUT: session id */
- int *lstio_ses_keyp; /* OUT: local key */
+ /* OUT: session id */
+ lst_sid_t *lstio_ses_idp;
+ /* OUT: local key */
+ int *lstio_ses_keyp;
/** OUT: session features */
- unsigned *lstio_ses_featp;
- lstcon_ndlist_ent_t *lstio_ses_ndinfo; /* OUT: */
- int lstio_ses_nmlen; /* IN: name length */
- char *lstio_ses_namep; /* OUT: session name */
+ unsigned *lstio_ses_featp;
+ /* OUT: */
+ lstcon_ndlist_ent_t *lstio_ses_ndinfo;
+ /* IN: name length */
+ int lstio_ses_nmlen;
+ /* OUT: session name */
+ char *lstio_ses_namep;
} lstio_session_info_args_t;
/* delete a session */
typedef struct {
- int lstio_ses_key; /* IN: session key */
+ /* IN: session key */
+ int lstio_ses_key;
} lstio_session_end_args_t;
-#define LST_OPC_SESSION 1
-#define LST_OPC_GROUP 2
-#define LST_OPC_NODES 3
+#define LST_OPC_SESSION 1
+#define LST_OPC_GROUP 2
+#define LST_OPC_NODES 3
#define LST_OPC_BATCHCLI 4
#define LST_OPC_BATCHSRV 5
typedef struct {
- int lstio_dbg_key; /* IN: session key */
- int lstio_dbg_type; /* IN: debug sessin|batch|group|nodes list */
- int lstio_dbg_flags; /* IN: reserved debug flags */
- int lstio_dbg_timeout; /* IN: timeout of debug */
-
- int lstio_dbg_nmlen; /* IN: len of name */
- char *lstio_dbg_namep; /* IN: name of group|batch */
- int lstio_dbg_count; /* IN: # of test nodes to debug */
- lnet_process_id_t *lstio_dbg_idsp; /* IN: id of test nodes */
- struct list_head *lstio_dbg_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_dbg_key;
+ /* IN: debug sessin|batch|group|nodes list */
+ int lstio_dbg_type;
+ /* IN: reserved debug flags */
+ int lstio_dbg_flags;
+ /* IN: timeout of debug */
+ int lstio_dbg_timeout;
+ /* IN: len of name */
+ int lstio_dbg_nmlen;
+ /* IN: name of group|batch */
+ char *lstio_dbg_namep;
+ /* IN: # of test nodes to debug */
+ int lstio_dbg_count;
+ /* IN: id of test nodes */
+ lnet_process_id_t *lstio_dbg_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_dbg_resultp;
} lstio_debug_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
} lstio_group_add_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen; /* IN: name length */
+ /* IN: group name */
+ char *lstio_grp_namep;
} lstio_group_del_args_t;
-#define LST_GROUP_CLEAN 1 /* remove inactive nodes in the group */
-#define LST_GROUP_REFRESH 2 /* refresh inactive nodes in the group */
-#define LST_GROUP_RMND 3 /* delete nodes from the group */
-
-typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_opc; /* IN: OPC */
- int lstio_grp_args; /* IN: arguments */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
- int lstio_grp_count; /* IN: # of nodes id */
- lnet_process_id_t *lstio_grp_idsp; /* IN: array of nodes */
- struct list_head *lstio_grp_resultp; /* OUT: list head of result buffer */
+#define LST_GROUP_CLEAN 1 /* remove inactive nodes in the group */
+#define LST_GROUP_REFRESH 2 /* refresh inactive nodes in the group */
+#define LST_GROUP_RMND 3 /* delete nodes from the group */
+
+typedef struct {
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: OPC */
+ int lstio_grp_opc;
+ /* IN: arguments */
+ int lstio_grp_args;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
+ /* IN: # of nodes id */
+ int lstio_grp_count;
+ /* IN: array of nodes */
+ lnet_process_id_t *lstio_grp_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_grp_resultp;
} lstio_group_update_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name length */
- char *lstio_grp_namep; /* IN: group name */
- int lstio_grp_count; /* IN: # of nodes */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name length */
+ int lstio_grp_nmlen;
+ /* IN: group name */
+ char *lstio_grp_namep;
+ /* IN: # of nodes */
+ int lstio_grp_count;
/** OUT: session features */
- unsigned *lstio_grp_featp;
- lnet_process_id_t *lstio_grp_idsp; /* IN: nodes */
- struct list_head *lstio_grp_resultp; /* OUT: list head of result buffer */
+ unsigned *lstio_grp_featp;
+ /* IN: nodes */
+ lnet_process_id_t *lstio_grp_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_grp_resultp;
} lstio_group_nodes_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_idx; /* IN: group idx */
- int lstio_grp_nmlen; /* IN: name len */
- char *lstio_grp_namep; /* OUT: name */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: group idx */
+ int lstio_grp_idx;
+ /* IN: name len */
+ int lstio_grp_nmlen;
+ /* OUT: name */
+ char *lstio_grp_namep;
} lstio_group_list_args_t;
typedef struct {
- int lstio_grp_key; /* IN: session key */
- int lstio_grp_nmlen; /* IN: name len */
- char *lstio_grp_namep; /* IN: name */
- lstcon_ndlist_ent_t *lstio_grp_entp; /* OUT: description of group */
-
- int *lstio_grp_idxp; /* IN/OUT: node index */
- int *lstio_grp_ndentp; /* IN/OUT: # of nodent */
- lstcon_node_ent_t *lstio_grp_dentsp; /* OUT: nodent array */
+ /* IN: session key */
+ int lstio_grp_key;
+ /* IN: name len */
+ int lstio_grp_nmlen;
+ /* IN: name */
+ char *lstio_grp_namep;
+ /* OUT: description of group */
+ lstcon_ndlist_ent_t *lstio_grp_entp;
+ /* IN/OUT: node index */
+ int *lstio_grp_idxp;
+ /* IN/OUT: # of nodent */
+ int *lstio_grp_ndentp;
+ /* OUT: nodent array */
+ lstcon_node_ent_t *lstio_grp_dentsp;
} lstio_group_info_args_t;
-#define LST_DEFAULT_BATCH "batch" /* default batch name */
+#define LST_DEFAULT_BATCH "batch" /* default batch name */
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_add_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key; /* IN: session key */
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_del_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_timeout; /* IN: timeout for the batch */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: timeout for the batch */
+ int lstio_bat_timeout;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_run_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_force; /* IN: abort unfinished test RPC */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: abort unfinished test RPC */
+ int lstio_bat_force;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_stop_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_testidx; /* IN: test index */
- int lstio_bat_client; /* IN: is test client? */
- int lstio_bat_timeout; /* IN: timeout for waiting */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
- struct list_head *lstio_bat_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: test index */
+ int lstio_bat_testidx;
+ /* IN: is test client? */
+ int lstio_bat_client;
+ /* IN: timeout for waiting */
+ int lstio_bat_timeout;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_bat_resultp;
} lstio_batch_query_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_idx; /* IN: index */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: batch name */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: index */
+ int lstio_bat_idx;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_bat_namep;
} lstio_batch_list_args_t;
typedef struct {
- int lstio_bat_key; /* IN: session key */
- int lstio_bat_nmlen; /* IN: name length */
- char *lstio_bat_namep; /* IN: name */
- int lstio_bat_server; /* IN: query server or not */
- int lstio_bat_testidx; /* IN: test index */
- lstcon_test_batch_ent_t *lstio_bat_entp; /* OUT: batch ent */
-
- int *lstio_bat_idxp; /* IN/OUT: index of node */
- int *lstio_bat_ndentp; /* IN/OUT: # of nodent */
- lstcon_node_ent_t *lstio_bat_dentsp; /* array of nodent */
+ /* IN: session key */
+ int lstio_bat_key;
+ /* IN: name length */
+ int lstio_bat_nmlen;
+ /* IN: name */
+ char *lstio_bat_namep;
+ /* IN: query server or not */
+ int lstio_bat_server;
+ /* IN: test index */
+ int lstio_bat_testidx;
+ /* OUT: batch ent */
+ lstcon_test_batch_ent_t *lstio_bat_entp;
+ /* IN/OUT: index of node */
+ int *lstio_bat_idxp;
+ /* IN/OUT: # of nodent */
+ int *lstio_bat_ndentp;
+ /* array of nodent */
+ lstcon_node_ent_t *lstio_bat_dentsp;
} lstio_batch_info_args_t;
/* add stat in session */
typedef struct {
- int lstio_sta_key; /* IN: session key */
- int lstio_sta_timeout; /* IN: timeout for stat request */
- int lstio_sta_nmlen; /* IN: group name length */
- char *lstio_sta_namep; /* IN: group name */
- int lstio_sta_count; /* IN: # of pid */
- lnet_process_id_t *lstio_sta_idsp; /* IN: pid */
- struct list_head *lstio_sta_resultp; /* OUT: list head of result buffer */
+ /* IN: session key */
+ int lstio_sta_key;
+ /* IN: timeout for stat request */
+ int lstio_sta_timeout;
+ /* IN: group name length */
+ int lstio_sta_nmlen;
+ /* IN: group name */
+ char *lstio_sta_namep;
+ /* IN: # of pid */
+ int lstio_sta_count;
+ /* IN: pid */
+ lnet_process_id_t *lstio_sta_idsp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_sta_resultp;
} lstio_stat_args_t;
typedef enum {
- LST_TEST_BULK = 1,
- LST_TEST_PING = 2
+ LST_TEST_BULK = 1,
+ LST_TEST_PING = 2
} lst_test_type_t;
/* create a test in a batch */
-#define LST_MAX_CONCUR 1024 /* Max concurrency of test */
-
-typedef struct {
- int lstio_tes_key; /* IN: session key */
- int lstio_tes_bat_nmlen; /* IN: batch name len */
- char *lstio_tes_bat_name; /* IN: batch name */
- int lstio_tes_type; /* IN: test type */
- int lstio_tes_oneside; /* IN: one sided test */
- int lstio_tes_loop; /* IN: loop count */
- int lstio_tes_concur; /* IN: concurrency */
-
- int lstio_tes_dist; /* IN: node distribution in destination groups */
- int lstio_tes_span; /* IN: node span in destination groups */
- int lstio_tes_sgrp_nmlen; /* IN: source group name length */
- char *lstio_tes_sgrp_name; /* IN: group name */
- int lstio_tes_dgrp_nmlen; /* IN: destination group name length */
- char *lstio_tes_dgrp_name; /* IN: group name */
-
- int lstio_tes_param_len; /* IN: param buffer len */
- void *lstio_tes_param; /* IN: parameter for specified test:
- lstio_bulk_param_t,
- lstio_ping_param_t,
- ... more */
- int *lstio_tes_retp; /* OUT: private returned value */
- struct list_head *lstio_tes_resultp; /* OUT: list head of result buffer */
+#define LST_MAX_CONCUR 1024 /* Max concurrency of test */
+
+typedef struct {
+ /* IN: session key */
+ int lstio_tes_key;
+ /* IN: batch name len */
+ int lstio_tes_bat_nmlen;
+ /* IN: batch name */
+ char *lstio_tes_bat_name;
+ /* IN: test type */
+ int lstio_tes_type;
+ /* IN: one sided test */
+ int lstio_tes_oneside;
+ /* IN: loop count */
+ int lstio_tes_loop;
+ /* IN: concurrency */
+ int lstio_tes_concur;
+ /* IN: node distribution in destination groups */
+ int lstio_tes_dist;
+ /* IN: node span in destination groups */
+ int lstio_tes_span;
+ /* IN: source group name length */
+ int lstio_tes_sgrp_nmlen;
+ /* IN: group name */
+ char *lstio_tes_sgrp_name;
+ /* IN: destination group name length */
+ int lstio_tes_dgrp_nmlen;
+ /* IN: group name */
+ char *lstio_tes_dgrp_name;
+ /* IN: param buffer len */
+ int lstio_tes_param_len;
+ /* IN: parameter for specified test:
+ * lstio_bulk_param_t,
+ * lstio_ping_param_t,
+ * ... more */
+ void *lstio_tes_param;
+ /* OUT: private returned value */
+ int *lstio_tes_retp;
+ /* OUT: list head of result buffer */
+ struct list_head *lstio_tes_resultp;
} lstio_test_args_t;
typedef enum {
- LST_BRW_READ = 1,
- LST_BRW_WRITE = 2
+ LST_BRW_READ = 1,
+ LST_BRW_WRITE = 2
} lst_brw_type_t;
typedef enum {
- LST_BRW_CHECK_NONE = 1,
- LST_BRW_CHECK_SIMPLE = 2,
- LST_BRW_CHECK_FULL = 3
+ LST_BRW_CHECK_NONE = 1,
+ LST_BRW_CHECK_SIMPLE = 2,
+ LST_BRW_CHECK_FULL = 3
} lst_brw_flags_t;
typedef struct {
- int blk_opc; /* bulk operation code */
- int blk_size; /* size (bytes) */
- int blk_time; /* time of running the test*/
- int blk_flags; /* reserved flags */
+ /* bulk operation code */
+ int blk_opc;
+ /* size (bytes) */
+ int blk_size; /* size (bytes) */
+ /* time of running the test*/
+ int blk_time;
+ /* reserved flags */
+ int blk_flags;
} lst_test_bulk_param_t;
typedef struct {
- int png_size; /* size of ping message */
- int png_time; /* time */
- int png_loop; /* loop */
- int png_flags; /* reserved flags */
+ /* size of ping message */
+ int png_size;
+ /* time */
+ int png_time;
+ /* loop */
+ int png_loop;
+ /* reserved flags */
+ int png_flags;
} lst_test_ping_param_t;
-/* more tests */
typedef struct {
__u32 errors;
__u32 rpcs_sent;
diff --git a/drivers/staging/lustre/include/linux/lnet/nidstr.h b/drivers/staging/lustre/include/linux/lnet/nidstr.h
new file mode 100644
index 0000000..14cca80
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/nidstr.h
@@ -0,0 +1,73 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.gnu.org/licenses/gpl-2.0.html
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, 2014, Intel Corporation.
+ */
+#ifndef _LNET_NIDSTRINGS_H
+#define _LNET_NIDSTRINGS_H
+#include "types.h"
+
+/**
+ * Lustre Network Driver types.
+ */
+enum {
+ /* Only add to these values (i.e. don't ever change or redefine them):
+ * network addresses depend on them... */
+ QSWLND = 1,
+ SOCKLND = 2,
+ GMLND = 3,
+ PTLLND = 4,
+ O2IBLND = 5,
+ CIBLND = 6,
+ OPENIBLND = 7,
+ IIBLND = 8,
+ LOLND = 9,
+ RALND = 10,
+ VIBLND = 11,
+ MXLND = 12,
+ GNILND = 13,
+ GNIIPLND = 14,
+};
+
+/* # of nidstrings */
+#define LNET_NIDSTR_COUNT 1024
+/* size of each one (see below for usage) */
+#define LNET_NIDSTR_SIZE 32
+
+int libcfs_isknown_lnd(int type);
+char *libcfs_lnd2modname(int type);
+char *libcfs_lnd2str(int type);
+int libcfs_str2lnd(const char *str);
+char *libcfs_net2str(__u32 net);
+char *libcfs_nid2str(lnet_nid_t nid);
+__u32 libcfs_str2net(const char *str);
+lnet_nid_t libcfs_str2nid(const char *str);
+int libcfs_str2anynid(lnet_nid_t *nid, const char *str);
+char *libcfs_id2str(lnet_process_id_t id);
+void cfs_free_nidlist(struct list_head *list);
+int cfs_parse_nidlist(char *str, int len, struct list_head *list);
+int cfs_match_nid(lnet_nid_t nid, struct list_head *list);
+
+#endif /* _LNET_NIDSTRINGS_H */
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd.h b/drivers/staging/lustre/include/linux/lnet/ptllnd.h
deleted file mode 100644
index c91d653..0000000
--- a/drivers/staging/lustre/include/linux/lnet/ptllnd.h
+++ /dev/null
@@ -1,93 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lnet/include/lnet/ptllnd.h
- *
- * Author: PJ Kirner <pjkirner@clusterfs.com>
- */
-
-/*
- * The PTLLND was designed to support Portals with
- * Lustre and non-lustre UNLINK semantics.
- * However for now the two targets are Cray Portals
- * on the XT3 and Lustre Portals (for testing) both
- * have Lustre UNLINK semantics, so this is defined
- * by default.
- */
-#define LUSTRE_PORTALS_UNLINK_SEMANTICS
-
-#ifdef _USING_LUSTRE_PORTALS_
-
-/* NIDs are 64-bits on Lustre Portals */
-#define FMT_NID "%llu"
-#define FMT_PID "%d"
-
-/* When using Lustre Portals Lustre completion semantics are imlicit*/
-#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS 0
-
-#else /* _USING_CRAY_PORTALS_ */
-
-/* NIDs are integers on Cray Portals */
-#define FMT_NID "%u"
-#define FMT_PID "%d"
-
-/* When using Cray Portals this is defined in the Cray Portals Header*/
-/*#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS */
-
-/* Can compare handles directly on Cray Portals */
-#define PtlHandleIsEqual(a, b) ((a) == (b))
-
-/* Different error types on Cray Portals*/
-#define ptl_err_t ptl_ni_fail_t
-
-/*
- * The Cray Portals has no maximum number of IOVs. The
- * maximum is limited only by memory and size of the
- * int parameters (2^31-1).
- * Lustre only really require that the underyling
- * implementation to support at least LNET_MAX_IOV,
- * so for Cray portals we can safely just use that
- * value here.
- *
- */
-#define PTL_MD_MAX_IOV LNET_MAX_IOV
-
-#endif
-
-#define FMT_PTLID "ptlid:"FMT_PID"-"FMT_NID
-
-/* Align incoming small request messages to an 8 byte boundary if this is
- * supported to avoid alignment issues on some architectures */
-#ifndef PTL_MD_LOCAL_ALIGN8
-# define PTL_MD_LOCAL_ALIGN8 0
-#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h b/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
deleted file mode 100644
index 808f37b..0000000
--- a/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
+++ /dev/null
@@ -1,119 +0,0 @@
-/*
- * GPL HEADER START
- *
- * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
- *
- * This program is free software; you can redistribute it and/or modify
- * it under the terms of the GNU General Public License version 2 only,
- * as published by the Free Software Foundation.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- * General Public License version 2 for more details (a copy is included
- * in the LICENSE file that accompanied this code).
- *
- * You should have received a copy of the GNU General Public License
- * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
- *
- * GPL HEADER END
- */
-/*
- * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
- * Use is subject to license terms.
- */
-/*
- * This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
- *
- * lnet/include/lnet/ptllnd_wire.h
- *
- * Author: PJ Kirner <pjkirner@clusterfs.com>
- */
-
-/* Minimum buffer size that any peer will post to receive ptllnd messages */
-#define PTLLND_MIN_BUFFER_SIZE 256
-
-/************************************************************************
- * Tunable defaults that {u,k}lnds/ptllnd should have in common.
- */
-
-#define PTLLND_PORTAL 9 /* The same portal PTLPRC used when talking to cray portals */
-#define PTLLND_PID 9 /* The Portals PID */
-#define PTLLND_PEERCREDITS 8 /* concurrent sends to 1 peer */
-
-/* Default buffer size for kernel ptllnds (guaranteed eager) */
-#define PTLLND_MAX_KLND_MSG_SIZE 512
-
-/* Default buffer size for catamount ptllnds (not guaranteed eager) - large
- * enough to avoid RDMA for anything sent while control is not in liblustre */
-#define PTLLND_MAX_ULND_MSG_SIZE 512
-
-/************************************************************************
- * Portals LND Wire message format.
- * These are sent in sender's byte order (i.e. receiver flips).
- */
-
-#define PTL_RESERVED_MATCHBITS 0x100 /* below this value is reserved
- * above is for bulk data transfer */
-#define LNET_MSG_MATCHBITS 0 /* the value for the message channel */
-
-typedef struct {
- lnet_hdr_t kptlim_hdr; /* portals header */
- char kptlim_payload[0]; /* piggy-backed payload */
-} WIRE_ATTR kptl_immediate_msg_t;
-
-typedef struct {
- lnet_hdr_t kptlrm_hdr; /* portals header */
- __u64 kptlrm_matchbits; /* matchbits */
-} WIRE_ATTR kptl_rdma_msg_t;
-
-typedef struct {
- __u64 kptlhm_matchbits; /* matchbits */
- __u32 kptlhm_max_msg_size; /* max message size */
-} WIRE_ATTR kptl_hello_msg_t;
-
-typedef struct {
- /* First 2 fields fixed FOR ALL TIME */
- __u32 ptlm_magic; /* I'm a Portals LND message */
- __u16 ptlm_version; /* this is my version number */
- __u8 ptlm_type; /* the message type */
- __u8 ptlm_credits; /* returned credits */
- __u32 ptlm_nob; /* # bytes in whole message */
- __u32 ptlm_cksum; /* checksum (0 == no checksum) */
- __u64 ptlm_srcnid; /* sender's NID */
- __u64 ptlm_srcstamp; /* sender's incarnation */
- __u64 ptlm_dstnid; /* destination's NID */
- __u64 ptlm_dststamp; /* destination's incarnation */
- __u32 ptlm_srcpid; /* sender's PID */
- __u32 ptlm_dstpid; /* destination's PID */
-
- union {
- kptl_immediate_msg_t immediate;
- kptl_rdma_msg_t rdma;
- kptl_hello_msg_t hello;
- } WIRE_ATTR ptlm_u;
-
-} kptl_msg_t;
-
-/* kptl_msg_t::ptlm_credits is only a __u8 */
-#define PTLLND_MSG_MAX_CREDITS ((typeof(((kptl_msg_t *)0)->ptlm_credits)) - 1)
-
-#define PTLLND_MSG_MAGIC LNET_PROTO_PTL_MAGIC
-#define PTLLND_MSG_VERSION 0x04
-
-#define PTLLND_RDMA_OK 0x00
-#define PTLLND_RDMA_FAIL 0x01
-
-#define PTLLND_MSG_TYPE_INVALID 0x00
-#define PTLLND_MSG_TYPE_PUT 0x01
-#define PTLLND_MSG_TYPE_GET 0x02
-#define PTLLND_MSG_TYPE_IMMEDIATE 0x03 /* No bulk data xfer*/
-#define PTLLND_MSG_TYPE_NOOP 0x04
-#define PTLLND_MSG_TYPE_HELLO 0x05
-#define PTLLND_MSG_TYPE_NAK 0x06
diff --git a/drivers/staging/lustre/include/linux/lnet/socklnd.h b/drivers/staging/lustre/include/linux/lnet/socklnd.h
index 389038b..e62df8c 100644
--- a/drivers/staging/lustre/include/linux/lnet/socklnd.h
+++ b/drivers/staging/lustre/include/linux/lnet/socklnd.h
@@ -15,89 +15,92 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
+ *
+ * Copyright (c) 2012, 2014 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*
* lnet/include/lnet/socklnd.h
- *
- * #defines shared between socknal implementation and utilities
*/
#ifndef __LNET_LNET_SOCKLND_H__
#define __LNET_LNET_SOCKLND_H__
#include "types.h"
-#include "lib-types.h"
-#define SOCKLND_CONN_NONE (-1)
+#define SOCKLND_CONN_NONE (-1)
#define SOCKLND_CONN_ANY 0
-#define SOCKLND_CONN_CONTROL 1
-#define SOCKLND_CONN_BULK_IN 2
-#define SOCKLND_CONN_BULK_OUT 3
-#define SOCKLND_CONN_NTYPES 4
+#define SOCKLND_CONN_CONTROL 1
+#define SOCKLND_CONN_BULK_IN 2
+#define SOCKLND_CONN_BULK_OUT 3
+#define SOCKLND_CONN_NTYPES 4
#define SOCKLND_CONN_ACK SOCKLND_CONN_BULK_IN
typedef struct {
- __u32 kshm_magic; /* magic number of socklnd message */
- __u32 kshm_version; /* version of socklnd message */
- lnet_nid_t kshm_src_nid; /* sender's nid */
- lnet_nid_t kshm_dst_nid; /* destination nid */
- lnet_pid_t kshm_src_pid; /* sender's pid */
- lnet_pid_t kshm_dst_pid; /* destination pid */
- __u64 kshm_src_incarnation; /* sender's incarnation */
- __u64 kshm_dst_incarnation; /* destination's incarnation */
- __u32 kshm_ctype; /* connection type */
- __u32 kshm_nips; /* # IP addrs */
- __u32 kshm_ips[0]; /* IP addrs */
+ /* magic number of socklnd message */
+ __u32 kshm_magic;
+ /* version of socklnd message */
+ __u32 kshm_version;
+ /* sender's nid */
+ lnet_nid_t kshm_src_nid;
+ /* destination nid */
+ lnet_nid_t kshm_dst_nid;
+ /* sender's pid */
+ lnet_pid_t kshm_src_pid;
+ /* destination pid */
+ lnet_pid_t kshm_dst_pid;
+ /* sender's incarnation */
+ __u64 kshm_src_incarnation;
+ /* destination's incarnation */
+ __u64 kshm_dst_incarnation;
+ /* connection type */
+ __u32 kshm_ctype;
+ /* # IP addrs */
+ __u32 kshm_nips;
+ __u32 kshm_ips[0];
} WIRE_ATTR ksock_hello_msg_t;
typedef struct {
- lnet_hdr_t ksnm_hdr; /* lnet hdr */
-
- /*
- * ksnm_payload is removed because of winnt compiler's limitation:
- * zero-sized array can only be placed at the tail of [nested]
- * structure definitions. lnet payload will be stored just after
- * the body of structure ksock_lnet_msg_t
- */
+ /* lnet hdr */
+ lnet_hdr_t ksnm_hdr;
} WIRE_ATTR ksock_lnet_msg_t;
typedef struct {
- __u32 ksm_type; /* type of socklnd message */
- __u32 ksm_csum; /* checksum if != 0 */
- __u64 ksm_zc_cookies[2]; /* Zero-Copy request/ACK cookie */
+ /* type of socklnd message */
+ __u32 ksm_type;
+ /* checksum if != 0 */
+ __u32 ksm_csum;
+ /* Zero-Copy request/ACK cookie */
+ __u64 ksm_zc_cookies[2];
union {
- ksock_lnet_msg_t lnetmsg; /* lnet message, it's empty if it's NOOP */
+ /* lnet message, it's empty if it's NOOP */
+ ksock_lnet_msg_t lnetmsg;
} WIRE_ATTR ksm_u;
} WIRE_ATTR ksock_msg_t;
static inline void
socklnd_init_msg(ksock_msg_t *msg, int type)
{
- msg->ksm_csum = 0;
- msg->ksm_type = type;
- msg->ksm_zc_cookies[0] = msg->ksm_zc_cookies[1] = 0;
+ msg->ksm_csum = 0;
+ msg->ksm_type = type;
+ msg->ksm_zc_cookies[0] = msg->ksm_zc_cookies[1] = 0;
}
-#define KSOCK_MSG_NOOP 0xc0 /* ksm_u empty */
-#define KSOCK_MSG_LNET 0xc1 /* lnet msg */
+#define KSOCK_MSG_NOOP 0xc0 /* ksm_u empty */
+#define KSOCK_MSG_LNET 0xc1 /* lnet msg */
/* We need to know this number to parse hello msg from ksocklnd in
* other LND (usocklnd, for example) */
-#define KSOCK_PROTO_V2 2
-#define KSOCK_PROTO_V3 3
+#define KSOCK_PROTO_V2 2
+#define KSOCK_PROTO_V3 3
#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/types.h b/drivers/staging/lustre/include/linux/lnet/types.h
index 68d8139..53997ae 100644
--- a/drivers/staging/lustre/include/linux/lnet/types.h
+++ b/drivers/staging/lustre/include/linux/lnet/types.h
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,11 +23,11 @@
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2014 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#ifndef __LNET_TYPES_H__
@@ -48,7 +44,7 @@
/** Portal reserved for LNet's own use.
* \see lustre/include/lustre/lustre_idl.h for Lustre portal assignments.
*/
-#define LNET_RESERVED_PORTAL 0
+#define LNET_RESERVED_PORTAL 0
/**
* Address of an end-point in an LNet network.
@@ -74,8 +70,179 @@ typedef __u32 lnet_pid_t;
#define LNET_PID_RESERVED 0xf0000000 /* reserved bits in PID */
#define LNET_PID_USERFLAG 0x80000000 /* set in userspace peers */
+#define LNET_PID_LUSTRE 12345
+
+#define LNET_TIME_FOREVER (-1)
+
+/* how an LNET NID encodes net:address */
+/** extract the address part of an lnet_nid_t */
+
+static inline __u32 LNET_NIDADDR(lnet_nid_t nid)
+{
+ return nid & 0xffffffff;
+}
+
+static inline __u32 LNET_NIDNET(lnet_nid_t nid)
+{
+ return (nid >> 32) & 0xffffffff;
+}
+
+static inline lnet_nid_t LNET_MKNID(__u32 net, __u32 addr)
+{
+ return (((__u64)net) << 32) | addr;
+}
+
+static inline __u32 LNET_NETNUM(__u32 net)
+{
+ return net & 0xffff;
+}
+
+static inline __u32 LNET_NETTYP(__u32 net)
+{
+ return (net >> 16) & 0xffff;
+}
+
+static inline __u32 LNET_MKNET(__u32 type, __u32 num)
+{
+ return (type << 16) | num;
+}
+
+#define WIRE_ATTR __packed
+
+/* Packed version of lnet_process_id_t to transfer via network */
+typedef struct {
+ /* node id / process id */
+ lnet_nid_t nid;
+ lnet_pid_t pid;
+} WIRE_ATTR lnet_process_id_packed_t;
-#define LNET_TIME_FOREVER (-1)
+/* The wire handle's interface cookie only matches one network interface in
+ * one epoch (i.e. new cookie when the interface restarts or the node
+ * reboots). The object cookie only matches one object on that interface
+ * during that object's lifetime (i.e. no cookie re-use). */
+typedef struct {
+ __u64 wh_interface_cookie;
+ __u64 wh_object_cookie;
+} WIRE_ATTR lnet_handle_wire_t;
+
+typedef enum {
+ LNET_MSG_ACK = 0,
+ LNET_MSG_PUT,
+ LNET_MSG_GET,
+ LNET_MSG_REPLY,
+ LNET_MSG_HELLO,
+} lnet_msg_type_t;
+
+/* The variant fields of the portals message header are aligned on an 8
+ * byte boundary in the message header. Note that all types used in these
+ * wire structs MUST be fixed size and the smaller types are placed at the
+ * end. */
+typedef struct lnet_ack {
+ lnet_handle_wire_t dst_wmd;
+ __u64 match_bits;
+ __u32 mlength;
+} WIRE_ATTR lnet_ack_t;
+
+typedef struct lnet_put {
+ lnet_handle_wire_t ack_wmd;
+ __u64 match_bits;
+ __u64 hdr_data;
+ __u32 ptl_index;
+ __u32 offset;
+} WIRE_ATTR lnet_put_t;
+
+typedef struct lnet_get {
+ lnet_handle_wire_t return_wmd;
+ __u64 match_bits;
+ __u32 ptl_index;
+ __u32 src_offset;
+ __u32 sink_length;
+} WIRE_ATTR lnet_get_t;
+
+typedef struct lnet_reply {
+ lnet_handle_wire_t dst_wmd;
+} WIRE_ATTR lnet_reply_t;
+
+typedef struct lnet_hello {
+ __u64 incarnation;
+ __u32 type;
+} WIRE_ATTR lnet_hello_t;
+
+typedef struct {
+ lnet_nid_t dest_nid;
+ lnet_nid_t src_nid;
+ lnet_pid_t dest_pid;
+ lnet_pid_t src_pid;
+ __u32 type; /* lnet_msg_type_t */
+ __u32 payload_length; /* payload data to follow */
+ /*<------__u64 aligned------->*/
+ union {
+ lnet_ack_t ack;
+ lnet_put_t put;
+ lnet_get_t get;
+ lnet_reply_t reply;
+ lnet_hello_t hello;
+ } msg;
+} WIRE_ATTR lnet_hdr_t;
+
+/* A HELLO message contains a magic number and protocol version
+ * code in the header's dest_nid, the peer's NID in the src_nid, and
+ * LNET_MSG_HELLO in the type field. All other common fields are zero
+ * (including payload_size; i.e. no payload).
+ * This is for use by byte-stream LNDs (e.g. TCP/IP) to check the peer is
+ * running the same protocol and to find out its NID. These LNDs should
+ * exchange HELLO messages when a connection is first established. Individual
+ * LNDs can put whatever else they fancy in lnet_hdr_t::msg.
+ */
+typedef struct {
+ __u32 magic; /* LNET_PROTO_TCP_MAGIC */
+ __u16 version_major; /* increment on incompatible change */
+ __u16 version_minor; /* increment on compatible change */
+} WIRE_ATTR lnet_magicversion_t;
+
+/* PROTO MAGIC for LNDs */
+#define LNET_PROTO_IB_MAGIC 0x0be91b91
+#define LNET_PROTO_GNI_MAGIC 0xb00fbabe /* ask Kim */
+#define LNET_PROTO_TCP_MAGIC 0xeebc0ded
+#define LNET_PROTO_ACCEPTOR_MAGIC 0xacce7100
+#define LNET_PROTO_PING_MAGIC 0x70696E67 /* 'ping' */
+
+/* Placeholder for a future "unified" protocol across all LNDs */
+/* Current LNDs that receive a request with this magic will respond with a
+ * "stub" reply using their current protocol */
+#define LNET_PROTO_MAGIC 0x45726963 /* ! */
+
+#define LNET_PROTO_TCP_VERSION_MAJOR 1
+#define LNET_PROTO_TCP_VERSION_MINOR 0
+
+/* Acceptor connection request */
+typedef struct {
+ __u32 acr_magic; /* PTL_ACCEPTOR_PROTO_MAGIC */
+ __u32 acr_version; /* protocol version */
+ __u64 acr_nid; /* target NID */
+} WIRE_ATTR lnet_acceptor_connreq_t;
+
+#define LNET_PROTO_ACCEPTOR_VERSION 1
+
+typedef struct lnet_counters {
+ __u32 msgs_alloc;
+ __u32 msgs_max;
+ __u32 errors;
+ __u32 send_count;
+ __u32 recv_count;
+ __u32 route_count;
+ __u32 drop_count;
+ __u64 send_length;
+ __u64 recv_length;
+ __u64 route_length;
+ __u64 drop_length;
+} WIRE_ATTR lnet_counters_t;
+
+#define LNET_NI_STATUS_UP 0x15aac0de
+#define LNET_NI_STATUS_DOWN 0xdeadface
+#define LNET_NI_STATUS_INVALID 0x00000000
+
+#define LNET_MAX_INTERFACES 16
/**
* Objects maintained by the LNet are accessed through handles. Handle types
@@ -195,8 +362,8 @@ typedef struct {
* one must start on page boundary, and all but the last must end on
* page boundary.
*/
- void *start;
- unsigned int length;
+ void *start;
+ unsigned int length;
/**
* Specifies the maximum number of operations that can be performed
* on the memory descriptor. An operation is any action that could
@@ -207,7 +374,7 @@ typedef struct {
* there is no bound on the number of operations that may be applied
* to a MD.
*/
- int threshold;
+ int threshold;
/**
* Specifies the largest incoming request that the memory descriptor
* should respond to. When the unused portion of a MD (length -
@@ -215,7 +382,7 @@ typedef struct {
* does not respond to further operations. This value is only used
* if the LNET_MD_MAX_SIZE option is set.
*/
- int max_size;
+ int max_size;
/**
* Specifies the behavior of the memory descriptor. A bitwise OR
* of the following values can be used:
@@ -252,14 +419,14 @@ typedef struct {
* region (i.e. sum of all fragment lengths) must not be less than
* \a max_size.
*/
- unsigned int options;
+ unsigned int options;
/**
* A user-specified value that is associated with the memory
* descriptor. The value does not need to be a pointer, but must fit
* in the space used by a pointer. This value is recorded in events
* associated with operations on this MD.
*/
- void *user_ptr;
+ void *user_ptr;
/**
* A handle for the event queue used to log the operations performed on
* the memory region. If this argument is a NULL handle (i.e. nullified
@@ -276,44 +443,33 @@ typedef struct {
#define LNET_MTU (1 << LNET_MTU_BITS)
/** limit on the number of fragments in discontiguous MDs */
-#define LNET_MAX_IOV 256
-
-/* Max payload size */
-# define LNET_MAX_PAYLOAD CONFIG_LNET_MAX_PAYLOAD
-# if (LNET_MAX_PAYLOAD < LNET_MTU)
-# error "LNET_MAX_PAYLOAD too small - error in configure --with-max-payload-mb"
-# else
-# if (LNET_MAX_PAYLOAD > (PAGE_SIZE * LNET_MAX_IOV))
-/* PAGE_SIZE is a constant: check with cpp! */
-# error "LNET_MAX_PAYLOAD too large - error in configure --with-max-payload-mb"
-# endif
-# endif
+#define LNET_MAX_IOV 256
/**
* Options for the MD structure. See lnet_md_t::options.
*/
-#define LNET_MD_OP_PUT (1 << 0)
+#define LNET_MD_OP_PUT (1 << 0)
/** See lnet_md_t::options. */
-#define LNET_MD_OP_GET (1 << 1)
+#define LNET_MD_OP_GET (1 << 1)
/** See lnet_md_t::options. */
#define LNET_MD_MANAGE_REMOTE (1 << 2)
-/* unused (1 << 3) */
+/* unused (1 << 3) */
/** See lnet_md_t::options. */
-#define LNET_MD_TRUNCATE (1 << 4)
+#define LNET_MD_TRUNCATE (1 << 4)
/** See lnet_md_t::options. */
-#define LNET_MD_ACK_DISABLE (1 << 5)
+#define LNET_MD_ACK_DISABLE (1 << 5)
/** See lnet_md_t::options. */
#define LNET_MD_IOVEC (1 << 6)
/** See lnet_md_t::options. */
-#define LNET_MD_MAX_SIZE (1 << 7)
+#define LNET_MD_MAX_SIZE (1 << 7)
/** See lnet_md_t::options. */
-#define LNET_MD_KIOV (1 << 8)
+#define LNET_MD_KIOV (1 << 8)
/* For compatibility with Cray Portals */
-#define LNET_MD_PHYS 0
+#define LNET_MD_PHYS 0
/** Infinite threshold on MD operations. See lnet_md_t::threshold */
-#define LNET_MD_THRESH_INF (-1)
+#define LNET_MD_THRESH_INF (-1)
/* NB lustre portals uses struct iovec internally! */
typedef struct iovec lnet_md_iovec_t;
@@ -323,15 +479,15 @@ typedef struct iovec lnet_md_iovec_t;
*/
typedef struct {
/** Pointer to the page where the fragment resides */
- struct page *kiov_page;
+ struct page *kiov_page;
/** Length in bytes of the fragment */
- unsigned int kiov_len;
+ unsigned int kiov_len;
/**
* Starting offset of the fragment within the page. Note that the
* end of the fragment must not pass the end of the page; i.e.,
* kiov_len + kiov_offset <= PAGE_CACHE_SIZE.
*/
- unsigned int kiov_offset;
+ unsigned int kiov_offset;
} lnet_kiov_t;
/** @} lnet_md */
@@ -379,7 +535,7 @@ typedef enum {
LNET_EVENT_UNLINK,
} lnet_event_kind_t;
-#define LNET_SEQ_BASETYPE long
+#define LNET_SEQ_BASETYPE long
typedef unsigned LNET_SEQ_BASETYPE lnet_seq_t;
#define LNET_SEQ_GT(a, b) (((signed LNET_SEQ_BASETYPE)((a) - (b))) > 0)
@@ -388,23 +544,23 @@ typedef unsigned LNET_SEQ_BASETYPE lnet_seq_t;
*/
typedef struct {
/** The identifier (nid, pid) of the target. */
- lnet_process_id_t target;
+ lnet_process_id_t target;
/** The identifier (nid, pid) of the initiator. */
- lnet_process_id_t initiator;
+ lnet_process_id_t initiator;
/**
* The NID of the immediate sender. If the request has been forwarded
* by routers, this is the NID of the last hop; otherwise it's the
* same as the initiator.
*/
- lnet_nid_t sender;
+ lnet_nid_t sender;
/** Indicates the type of the event. */
- lnet_event_kind_t type;
+ lnet_event_kind_t type;
/** The portal table index specified in the request */
- unsigned int pt_index;
+ unsigned int pt_index;
/** A copy of the match bits specified in the request. */
- __u64 match_bits;
+ __u64 match_bits;
/** The length (in bytes) specified in the request. */
- unsigned int rlength;
+ unsigned int rlength;
/**
* The length (in bytes) of the data that was manipulated by the
* operation. For truncated operations, the manipulated length will be
@@ -412,47 +568,47 @@ typedef struct {
* see lnet_md_t). For all other operations, the manipulated length
* will be the length of the requested operation, i.e. rlength.
*/
- unsigned int mlength;
+ unsigned int mlength;
/**
* The handle to the MD associated with the event. The handle may be
* invalid if the MD has been unlinked.
*/
- lnet_handle_md_t md_handle;
+ lnet_handle_md_t md_handle;
/**
* A snapshot of the state of the MD immediately after the event has
* been processed. In particular, the threshold field in md will
* reflect the value of the threshold after the operation occurred.
*/
- lnet_md_t md;
+ lnet_md_t md;
/**
* 64 bits of out-of-band user data. Only valid for LNET_EVENT_PUT.
* \see LNetPut
*/
- __u64 hdr_data;
+ __u64 hdr_data;
/**
* Indicates the completion status of the operation. It's 0 for
* successful operations, otherwise it's an error code.
*/
- int status;
+ int status;
/**
* Indicates whether the MD has been unlinked. Note that:
* - An event with unlinked set is the last event on the MD.
* - This field is also set for an explicit LNET_EVENT_UNLINK event.
* \see LNetMDUnlink
*/
- int unlinked;
+ int unlinked;
/**
* The displacement (in bytes) into the memory region that the
* operation used. The offset can be determined by the operation for
* a remote managed MD or by the local MD.
* \see lnet_md_t::options
*/
- unsigned int offset;
+ unsigned int offset;
/**
* The sequence number for this event. Sequence numbers are unique
* to each event.
*/
- volatile lnet_seq_t sequence;
+ volatile lnet_seq_t sequence;
} lnet_event_t;
/**
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index be4de56..8a3341e 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -1177,7 +1177,7 @@ void kiblnd_map_rx_descs(kib_conn_t *conn)
CDEBUG(D_NET, "rx %d: %p %#llx(%#llx)\n",
i, rx->rx_msg, rx->rx_msgaddr,
- lnet_page2phys(pg) + pg_off);
+ (__u64)(page_to_phys(pg) + pg_off));
pg_off += IBLND_MSG_SIZE;
LASSERT(pg_off <= PAGE_SIZE);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
index cd664d0..32b39a4 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.h
@@ -62,16 +62,15 @@
#define DEBUG_SUBSYSTEM S_LND
-#include "../../../include/linux/libcfs/libcfs.h"
-#include "../../../include/linux/lnet/lnet.h"
-#include "../../../include/linux/lnet/lib-lnet.h"
-#include "../../../include/linux/lnet/lnet-sysctl.h"
-
#include <rdma/rdma_cm.h>
#include <rdma/ib_cm.h>
#include <rdma/ib_verbs.h>
#include <rdma/ib_fmr_pool.h>
+#include "../../../include/linux/libcfs/libcfs.h"
+#include "../../../include/linux/lnet/lnet.h"
+#include "../../../include/linux/lnet/lib-lnet.h"
+
#define IBLND_PEER_HASH_SIZE 101 /* # peer lists */
/* # scheduler loops before reschedule */
#define IBLND_RESCHED 100
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index f3b4923..8798e49 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -34,7 +34,6 @@
#include "../../../include/linux/lnet/lnet.h"
#include "../../../include/linux/lnet/lib-lnet.h"
#include "../../../include/linux/lnet/socklnd.h"
-#include "../../../include/linux/lnet/lnet-sysctl.h"
#define SOCKNAL_PEER_HASH_SIZE 101 /* # peer lists */
#define SOCKNAL_RESCHED 100 /* # scheduler loops before reschedule */
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index a93f59c..df0f7d8 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -148,8 +148,8 @@ ksocknal_lib_send_kiov(ksock_conn_t *conn, ksock_tx_t *tx)
rc = sk->sk_prot->sendpage(sk, page,
offset, fragsize, msgflg);
} else {
- rc = cfs_tcp_sendpage(sk, page, offset, fragsize,
- msgflg);
+ rc = tcp_sendpage(sk, page, offset, fragsize,
+ msgflg);
}
} else {
#if SOCKNAL_SINGLE_FRAG_TX || !SOCKNAL_RISK_KMAP_DEADLOCK
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index bd97099..c44d3e0 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -243,8 +243,6 @@ lnet_accept(struct socket *sock, __u32 magic)
if (magic == le32_to_cpu(LNET_PROTO_TCP_MAGIC))
str = "'old' socknal/tcpnal";
- else if (lnet_accept_magic(magic, LNET_PROTO_RA_MAGIC))
- str = "'old' ranal";
else
str = "unrecognised";
diff --git a/drivers/staging/lustre/lnet/lnet/api-ni.c b/drivers/staging/lustre/lnet/lnet/api-ni.c
index 4a14e51..f9ae9e5 100644
--- a/drivers/staging/lustre/lnet/lnet/api-ni.c
+++ b/drivers/staging/lustre/lnet/lnet/api-ni.c
@@ -364,56 +364,6 @@ lnet_counters_reset(void)
}
EXPORT_SYMBOL(lnet_counters_reset);
-#ifdef LNET_USE_LIB_FREELIST
-
-int
-lnet_freelist_init(lnet_freelist_t *fl, int n, int size)
-{
- char *space;
-
- LASSERT(n > 0);
-
- size += offsetof(lnet_freeobj_t, fo_contents);
-
- LIBCFS_ALLOC(space, n * size);
- if (space == NULL)
- return -ENOMEM;
-
- INIT_LIST_HEAD(&fl->fl_list);
- fl->fl_objs = space;
- fl->fl_nobjs = n;
- fl->fl_objsize = size;
-
- do {
- memset(space, 0, size);
- list_add((struct list_head *)space, &fl->fl_list);
- space += size;
- } while (--n != 0);
-
- return 0;
-}
-
-void
-lnet_freelist_fini(lnet_freelist_t *fl)
-{
- struct list_head *el;
- int count;
-
- if (fl->fl_nobjs == 0)
- return;
-
- count = 0;
- for (el = fl->fl_list.next; el != &fl->fl_list; el = el->next)
- count++;
-
- LASSERT(count == fl->fl_nobjs);
-
- LIBCFS_FREE(fl->fl_objs, fl->fl_nobjs * fl->fl_objsize);
- memset(fl, 0, sizeof(*fl));
-}
-
-#endif /* LNET_USE_LIB_FREELIST */
-
static __u64
lnet_create_interface_cookie(void)
{
@@ -470,9 +420,6 @@ lnet_res_container_cleanup(struct lnet_res_container *rec)
count, lnet_res_type2str(rec->rec_type));
}
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_fini(&rec->rec_freelist);
-#endif
if (rec->rec_lh_hash != NULL) {
LIBCFS_FREE(rec->rec_lh_hash,
LNET_LH_HASH_SIZE * sizeof(rec->rec_lh_hash[0]));
@@ -483,8 +430,7 @@ lnet_res_container_cleanup(struct lnet_res_container *rec)
}
static int
-lnet_res_container_setup(struct lnet_res_container *rec,
- int cpt, int type, int objnum, int objsz)
+lnet_res_container_setup(struct lnet_res_container *rec, int cpt, int type)
{
int rc = 0;
int i;
@@ -494,12 +440,6 @@ lnet_res_container_setup(struct lnet_res_container *rec,
rec->rec_type = type;
INIT_LIST_HEAD(&rec->rec_active);
-#ifdef LNET_USE_LIB_FREELIST
- memset(&rec->rec_freelist, 0, sizeof(rec->rec_freelist));
- rc = lnet_freelist_init(&rec->rec_freelist, objnum, objsz);
- if (rc != 0)
- goto out;
-#endif
rec->rec_lh_cookie = (cpt << LNET_COOKIE_TYPE_BITS) | type;
/* Arbitrary choice of hash table size */
@@ -535,7 +475,7 @@ lnet_res_containers_destroy(struct lnet_res_container **recs)
}
static struct lnet_res_container **
-lnet_res_containers_create(int type, int objnum, int objsz)
+lnet_res_containers_create(int type)
{
struct lnet_res_container **recs;
struct lnet_res_container *rec;
@@ -550,7 +490,7 @@ lnet_res_containers_create(int type, int objnum, int objsz)
}
cfs_percpt_for_each(rec, i, recs) {
- rc = lnet_res_container_setup(rec, i, type, objnum, objsz);
+ rc = lnet_res_container_setup(rec, i, type);
if (rc != 0) {
lnet_res_containers_destroy(recs);
return NULL;
@@ -597,7 +537,6 @@ lnet_res_lh_initialize(struct lnet_res_container *rec, lnet_libhandle_t *lh)
list_add(&lh->lh_hash_chain, &rec->rec_lh_hash[hash]);
}
-
int lnet_unprepare(void);
static int
@@ -643,13 +582,11 @@ lnet_prepare(lnet_pid_t requested_pid)
goto failed;
rc = lnet_res_container_setup(&the_lnet.ln_eq_container, 0,
- LNET_COOKIE_TYPE_EQ, LNET_FL_MAX_EQS,
- sizeof(lnet_eq_t));
+ LNET_COOKIE_TYPE_EQ);
if (rc != 0)
goto failed;
- recs = lnet_res_containers_create(LNET_COOKIE_TYPE_ME, LNET_FL_MAX_MES,
- sizeof(lnet_me_t));
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_ME);
if (recs == NULL) {
rc = -ENOMEM;
goto failed;
@@ -657,8 +594,7 @@ lnet_prepare(lnet_pid_t requested_pid)
the_lnet.ln_me_containers = recs;
- recs = lnet_res_containers_create(LNET_COOKIE_TYPE_MD, LNET_FL_MAX_MDS,
- sizeof(lnet_libmd_t));
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_MD);
if (recs == NULL) {
rc = -ENOMEM;
goto failed;
@@ -1171,15 +1107,14 @@ lnet_startup_lndnis(void)
/**
* Initialize LNet library.
*
- * Only userspace program needs to call this function - it's automatically
- * called in the kernel at module loading time. Caller has to call LNetFini()
- * after a call to LNetInit(), if and only if the latter returned 0. It must
- * be called exactly once.
+ * Automatically called at module loading time. Caller has to call
+ * lnet_fini() after a call to lnet_init(), if and only if the latter
+ * returned 0. It must be called exactly once.
*
* \return 0 on success, and -ve on failures.
*/
int
-LNetInit(void)
+lnet_init(void)
{
int rc;
@@ -1232,19 +1167,16 @@ LNetInit(void)
lnet_register_lnd(&the_lolnd);
return 0;
}
-EXPORT_SYMBOL(LNetInit);
+EXPORT_SYMBOL(lnet_init);
/**
* Finalize LNet library.
*
- * Only userspace program needs to call this function. It can be called
- * at most once.
- *
- * \pre LNetInit() called with success.
- * \pre All LNet users called LNetNIFini() for matching LNetNIInit() calls.
+ * \pre lnet_init() called with success.
+ * \pre All LNet users called lnet_fini() for matching lnet_init() calls.
*/
void
-LNetFini(void)
+lnet_fini(void)
{
LASSERT(the_lnet.ln_init);
LASSERT(the_lnet.ln_refcount == 0);
@@ -1256,7 +1188,7 @@ LNetFini(void)
the_lnet.ln_init = 0;
}
-EXPORT_SYMBOL(LNetFini);
+EXPORT_SYMBOL(lnet_fini);
/**
* Set LNet PID and start LNet interfaces, routing, and forwarding.
@@ -1290,8 +1222,6 @@ LNetNIInit(lnet_pid_t requested_pid)
goto out;
}
- lnet_get_tunables();
-
if (requested_pid == LNET_PID_ANY) {
/* Don't instantiate LNET just for me */
rc = -ENETDOWN;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-eq.c b/drivers/staging/lustre/lnet/lnet/lib-eq.c
index 5470148..fabca5d 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-eq.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-eq.c
@@ -191,7 +191,7 @@ LNetEQFree(lnet_handle_eq_t eqh)
lnet_res_lh_invalidate(&eq->eq_lh);
list_del(&eq->eq_list);
- lnet_eq_free_locked(eq);
+ lnet_eq_free(eq);
out:
lnet_eq_wait_unlock();
lnet_res_unlock(LNET_LOCK_EX);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-md.c b/drivers/staging/lustre/lnet/lnet/lib-md.c
index 89d660f..57ede1b 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-md.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-md.c
@@ -82,7 +82,7 @@ lnet_md_unlink(lnet_libmd_t *md)
LASSERT(!list_empty(&md->md_list));
list_del_init(&md->md_list);
- lnet_md_free_locked(md);
+ lnet_md_free(md);
}
static int
@@ -320,7 +320,7 @@ LNetMDAttach(lnet_handle_me_t meh, lnet_md_t umd,
return 0;
failed:
- lnet_md_free_locked(md);
+ lnet_md_free(md);
lnet_res_unlock(cpt);
return rc;
@@ -381,7 +381,7 @@ LNetMDBind(lnet_md_t umd, lnet_unlink_t unlink, lnet_handle_md_t *handle)
return 0;
failed:
- lnet_md_free_locked(md);
+ lnet_md_free(md);
lnet_res_unlock(cpt);
return rc;
diff --git a/drivers/staging/lustre/lnet/lnet/lib-me.c b/drivers/staging/lustre/lnet/lnet/lib-me.c
index a3f9292..823df95 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-me.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-me.c
@@ -172,7 +172,7 @@ LNetMEInsert(lnet_handle_me_t current_meh,
current_me = lnet_handle2me(¤t_meh);
if (current_me == NULL) {
- lnet_me_free_locked(new_me);
+ lnet_me_free(new_me);
lnet_res_unlock(cpt);
return -ENOENT;
@@ -183,7 +183,7 @@ LNetMEInsert(lnet_handle_me_t current_meh,
ptl = the_lnet.ln_portals[current_me->me_portal];
if (lnet_ptl_is_unique(ptl)) {
/* nosense to insertion on unique portal */
- lnet_me_free_locked(new_me);
+ lnet_me_free(new_me);
lnet_res_unlock(cpt);
return -EPERM;
}
@@ -276,23 +276,5 @@ lnet_me_unlink(lnet_me_t *me)
}
lnet_res_lh_invalidate(&me->me_lh);
- lnet_me_free_locked(me);
+ lnet_me_free(me);
}
-
-#if 0
-static void
-lib_me_dump(lnet_me_t *me)
-{
- CWARN("Match Entry %p (%#llx)\n", me,
- me->me_lh.lh_cookie);
-
- CWARN("\tMatch/Ignore\t= %016lx / %016lx\n",
- me->me_match_bits, me->me_ignore_bits);
-
- CWARN("\tMD\t= %p\n", me->md);
- CWARN("\tprev\t= %p\n",
- list_entry(me->me_list.prev, lnet_me_t, me_list));
- CWARN("\tnext\t= %p\n",
- list_entry(me->me_list.next, lnet_me_t, me_list));
-}
-#endif
diff --git a/drivers/staging/lustre/lnet/lnet/lib-msg.c b/drivers/staging/lustre/lnet/lnet/lib-msg.c
index a46ccbf..41a5900 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-msg.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-msg.c
@@ -427,7 +427,7 @@ lnet_complete_msg_locked(lnet_msg_t *msg, int cpt)
}
lnet_msg_decommit(msg, cpt, status);
- lnet_msg_free_locked(msg);
+ lnet_msg_free(msg);
return 0;
}
@@ -559,35 +559,17 @@ lnet_msg_container_cleanup(struct lnet_msg_container *container)
sizeof(*container->msc_finalizers));
container->msc_finalizers = NULL;
}
-#ifdef LNET_USE_LIB_FREELIST
- lnet_freelist_fini(&container->msc_freelist);
-#endif
container->msc_init = 0;
}
int
lnet_msg_container_setup(struct lnet_msg_container *container, int cpt)
{
- int rc;
-
container->msc_init = 1;
INIT_LIST_HEAD(&container->msc_active);
INIT_LIST_HEAD(&container->msc_finalizing);
-#ifdef LNET_USE_LIB_FREELIST
- memset(&container->msc_freelist, 0, sizeof(lnet_freelist_t));
-
- rc = lnet_freelist_init(&container->msc_freelist,
- LNET_FL_MAX_MSGS, sizeof(lnet_msg_t));
- if (rc != 0) {
- CERROR("Failed to init freelist for message container\n");
- lnet_msg_container_cleanup(container);
- return rc;
- }
-#else
- rc = 0;
-#endif
/* number of CPUs */
container->msc_nfinalizers = cfs_cpt_weight(lnet_cpt_table(), cpt);
@@ -601,7 +583,7 @@ lnet_msg_container_setup(struct lnet_msg_container *container, int cpt)
return -ENOMEM;
}
- return rc;
+ return 0;
}
void
diff --git a/drivers/staging/lustre/lnet/lnet/module.c b/drivers/staging/lustre/lnet/lnet/module.c
index 72b7fbc..bc74738 100644
--- a/drivers/staging/lustre/lnet/lnet/module.c
+++ b/drivers/staging/lustre/lnet/lnet/module.c
@@ -117,7 +117,7 @@ init_lnet(void)
mutex_init(&lnet_config_mutex);
- rc = LNetInit();
+ rc = lnet_init();
if (rc != 0) {
CERROR("LNetInit: error %d\n", rc);
return rc;
@@ -143,7 +143,7 @@ fini_lnet(void)
rc = libcfs_deregister_ioctl(&lnet_ioctl_handler);
LASSERT(rc == 0);
- LNetFini();
+ lnet_fini();
}
MODULE_AUTHOR("Peter J. Braam <braam@clusterfs.com>");
diff --git a/drivers/staging/lustre/lnet/lnet/router.c b/drivers/staging/lustre/lnet/lnet/router.c
index 8510bae..6acaaae 100644
--- a/drivers/staging/lustre/lnet/lnet/router.c
+++ b/drivers/staging/lustre/lnet/lnet/router.c
@@ -24,8 +24,6 @@
#define DEBUG_SUBSYSTEM S_LNET
#include "../../include/linux/lnet/lib-lnet.h"
-#if defined(LNET_ROUTER)
-
#define LNET_NRB_TINY_MIN 512 /* min value for each CPT */
#define LNET_NRB_TINY (LNET_NRB_TINY_MIN * 4)
#define LNET_NRB_SMALL_MIN 4096 /* min value for each CPT */
@@ -70,15 +68,6 @@ lnet_peer_buffer_credits(lnet_ni_t *ni)
/* forward ref's */
static int lnet_router_checker(void *);
-#else
-
-int
-lnet_peer_buffer_credits(lnet_ni_t *ni)
-{
- return 0;
-}
-
-#endif
static int check_routers_before_use;
module_param(check_routers_before_use, int, 0444);
@@ -1163,9 +1152,6 @@ lnet_prune_rc_data(int wait_unlink)
lnet_net_unlock(LNET_LOCK_EX);
}
-
-#if defined(LNET_ROUTER)
-
static int
lnet_router_checker(void *arg)
{
@@ -1573,134 +1559,3 @@ lnet_notify(lnet_ni_t *ni, lnet_nid_t nid, int alive, unsigned long when)
return 0;
}
EXPORT_SYMBOL(lnet_notify);
-
-void
-lnet_get_tunables(void)
-{
-}
-
-#else
-
-int
-lnet_notify(lnet_ni_t *ni, lnet_nid_t nid, int alive, unsigned long when)
-{
- return -EOPNOTSUPP;
-}
-
-void
-lnet_router_checker(void)
-{
- static time_t last;
- static int running;
-
- time_t now = get_seconds();
- int interval = now - last;
- int rc;
- __u64 version;
- lnet_peer_t *rtr;
-
- /* It's no use to call me again within a sec - all intervals and
- * timeouts are measured in seconds */
- if (last != 0 && interval < 2)
- return;
-
- if (last != 0 &&
- interval > max(live_router_check_interval,
- dead_router_check_interval))
- CNETERR("Checker(%d/%d) not called for %d seconds\n",
- live_router_check_interval, dead_router_check_interval,
- interval);
-
- LASSERT(LNET_CPT_NUMBER == 1);
-
- lnet_net_lock(0);
- LASSERT(!running); /* recursion check */
- running = 1;
- lnet_net_unlock(0);
-
- last = now;
-
- if (the_lnet.ln_rc_state == LNET_RC_STATE_STOPPING)
- lnet_prune_rc_data(0); /* unlink all rcd and nowait */
-
- /* consume all pending events */
- while (1) {
- int i;
- lnet_event_t ev;
-
- /* NB ln_rc_eqh must be the 1st in 'eventqs' otherwise the
- * recursion breaker in LNetEQPoll would fail */
- rc = LNetEQPoll(&the_lnet.ln_rc_eqh, 1, 0, &ev, &i);
- if (rc == 0) /* no event pending */
- break;
-
- /* NB a lost SENT prevents me from pinging a router again */
- if (rc == -EOVERFLOW) {
- CERROR("Dropped an event!!!\n");
- abort();
- }
-
- LASSERT(rc == 1);
-
- lnet_router_checker_event(&ev);
- }
-
- if (the_lnet.ln_rc_state == LNET_RC_STATE_STOPPING) {
- lnet_prune_rc_data(1); /* release rcd */
- the_lnet.ln_rc_state = LNET_RC_STATE_SHUTDOWN;
- running = 0;
- return;
- }
-
- LASSERT(the_lnet.ln_rc_state == LNET_RC_STATE_RUNNING);
-
- lnet_net_lock(0);
-
- version = the_lnet.ln_routers_version;
- list_for_each_entry(rtr, &the_lnet.ln_routers, lp_rtr_list) {
- lnet_ping_router_locked(rtr);
- LASSERT(version == the_lnet.ln_routers_version);
- }
-
- lnet_net_unlock(0);
-
- running = 0; /* lock only needed for the recursion check */
-}
-
-/* NB lnet_peers_start_down depends on me,
- * so must be called before any peer creation */
-void
-lnet_get_tunables(void)
-{
- char *s;
-
- s = getenv("LNET_ROUTER_PING_TIMEOUT");
- if (s != NULL)
- router_ping_timeout = atoi(s);
-
- s = getenv("LNET_LIVE_ROUTER_CHECK_INTERVAL");
- if (s != NULL)
- live_router_check_interval = atoi(s);
-
- s = getenv("LNET_DEAD_ROUTER_CHECK_INTERVAL");
- if (s != NULL)
- dead_router_check_interval = atoi(s);
-
- /* This replaces old lnd_notify mechanism */
- check_routers_before_use = 1;
- if (dead_router_check_interval <= 0)
- dead_router_check_interval = 30;
-}
-
-void
-lnet_rtrpools_free(void)
-{
-}
-
-int
-lnet_rtrpools_alloc(int im_a_arouter)
-{
- return 0;
-}
-
-#endif
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index e2805bd..b32bc54 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -56,8 +56,8 @@
*/
#include "../../include/linux/libcfs/libcfs.h"
-// #include <obd.h>
-#include "../../include/linux/lnet/lnet.h"
+#include "../../include/linux/lnet/nidstr.h"
+#include "../../include/linux/lnet/api.h"
#include "lustre/lustre_idl.h"
#include "lustre_ha.h"
#include "lustre_sec.h"
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
2015-05-21 22:46 ` James Simmons
@ 2015-05-22 13:00 ` Dan Carpenter
-1 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 13:00 UTC (permalink / raw)
To: lustre-devel
This patch seems fine but it would also be better if it were broken up
into easy to review patches.
[patch 1] delete LNET_USE_LIB_FREELIST code
This is theortetically what this patch does except it does tons
of other things as well.
Patches which delete whole functions, #defines or from #ifdef to #endif
are normally easy to review. When there are + lines it is hard.
[patch 2] move things around (i guess)?
[patch 3] white space fixes
[patch 4] get rid of lnet_eq_free_locked() and macro friends.
[patch 5] rename CamelCase functions
[patch 6] clean up lnet_msg_container_setup()
> typedef struct lnet_peer {
> - struct list_head lp_hashlist; /* chain on peer hash */
> - struct list_head lp_txq; /* messages blocking for tx credits */
> - struct list_head lp_rtrq; /* messages blocking for router credits */
> - struct list_head lp_rtr_list; /* chain on router list */
> - int lp_txcredits; /* # tx credits available */
> - int lp_mintxcredits; /* low water mark */
> - int lp_rtrcredits; /* # router credits */
> - int lp_minrtrcredits; /* low water mark */
> - unsigned int lp_alive:1; /* alive/dead? */
> - unsigned int lp_notify:1; /* notification outstanding? */
> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
> - unsigned int lp_notifying:1; /* some thread is handling notification */
> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
> - int lp_alive_count; /* # times router went dead<->alive */
> - long lp_txqnob; /* bytes queued for sending */
> - unsigned long lp_timestamp; /* time of last aliveness news */
> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
> - unsigned long lp_last_alive; /* when I was last alive */
> - unsigned long lp_last_query; /* when lp_ni was queried last time */
> - lnet_ni_t *lp_ni; /* interface peer is on */
> - lnet_nid_t lp_nid; /* peer's NID */
> - int lp_refcount; /* # refs */
> - int lp_cpt; /* CPT this peer attached on */
> + /* chain on peer hash */
> + struct list_head lp_hashlist;
> + /* messages blocking for tx credits */
> + struct list_head lp_txq;
> + /* messages blocking for router credits */
> + struct list_head lp_rtrq;
> + /* chain on router list */
> + struct list_head lp_rtr_list;
> + /* # tx credits available */
> + int lp_txcredits;
> + /* low water mark */
> + int lp_mintxcredits;
> + /* # router credits */
> + int lp_rtrcredits;
> + /* low water mark */
> + int lp_minrtrcredits;
> + /* alive/dead? */
> + unsigned int lp_alive:1;
> + /* notification outstanding? */
> + unsigned int lp_notify:1;
> + /* outstanding notification for LND? */
> + unsigned int lp_notifylnd:1;
> + /* some thread is handling notification */
> + unsigned int lp_notifying:1;
> + /* SEND event outstanding from ping */
> + unsigned int lp_ping_notsent;
> + /* # times router went dead<->alive */
> + int lp_alive_count;
> + /* bytes queued for sending */
> + long lp_txqnob;
> + /* time of last aliveness news */
> + unsigned long lp_timestamp;
> + /* time of last ping attempt */
> + unsigned long lp_ping_timestamp;
> + /* != 0 if ping reply expected */
> + unsigned long lp_ping_deadline;
> + /* when I was last alive */
> + unsigned long lp_last_alive;
> + /* when lp_ni was queried last time */
> + unsigned long lp_last_query;
> + /* interface peer is on */
> + lnet_ni_t *lp_ni;
> + /* peer's NID */
> + lnet_nid_t lp_nid;
> + /* # refs */
> + int lp_refcount;
> + /* CPT this peer attached on */
> + int lp_cpt;
This new block of declarations is uglier than the original. Don't make
things uglier.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* Re: [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
@ 2015-05-22 13:00 ` Dan Carpenter
0 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-22 13:00 UTC (permalink / raw)
To: James Simmons
Cc: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger,
HPDD-discuss, James Simmons, James Simmons,
Linux Kernel Mailing List, lustre-devel
This patch seems fine but it would also be better if it were broken up
into easy to review patches.
[patch 1] delete LNET_USE_LIB_FREELIST code
This is theortetically what this patch does except it does tons
of other things as well.
Patches which delete whole functions, #defines or from #ifdef to #endif
are normally easy to review. When there are + lines it is hard.
[patch 2] move things around (i guess)?
[patch 3] white space fixes
[patch 4] get rid of lnet_eq_free_locked() and macro friends.
[patch 5] rename CamelCase functions
[patch 6] clean up lnet_msg_container_setup()
> typedef struct lnet_peer {
> - struct list_head lp_hashlist; /* chain on peer hash */
> - struct list_head lp_txq; /* messages blocking for tx credits */
> - struct list_head lp_rtrq; /* messages blocking for router credits */
> - struct list_head lp_rtr_list; /* chain on router list */
> - int lp_txcredits; /* # tx credits available */
> - int lp_mintxcredits; /* low water mark */
> - int lp_rtrcredits; /* # router credits */
> - int lp_minrtrcredits; /* low water mark */
> - unsigned int lp_alive:1; /* alive/dead? */
> - unsigned int lp_notify:1; /* notification outstanding? */
> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
> - unsigned int lp_notifying:1; /* some thread is handling notification */
> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
> - int lp_alive_count; /* # times router went dead<->alive */
> - long lp_txqnob; /* bytes queued for sending */
> - unsigned long lp_timestamp; /* time of last aliveness news */
> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
> - unsigned long lp_last_alive; /* when I was last alive */
> - unsigned long lp_last_query; /* when lp_ni was queried last time */
> - lnet_ni_t *lp_ni; /* interface peer is on */
> - lnet_nid_t lp_nid; /* peer's NID */
> - int lp_refcount; /* # refs */
> - int lp_cpt; /* CPT this peer attached on */
> + /* chain on peer hash */
> + struct list_head lp_hashlist;
> + /* messages blocking for tx credits */
> + struct list_head lp_txq;
> + /* messages blocking for router credits */
> + struct list_head lp_rtrq;
> + /* chain on router list */
> + struct list_head lp_rtr_list;
> + /* # tx credits available */
> + int lp_txcredits;
> + /* low water mark */
> + int lp_mintxcredits;
> + /* # router credits */
> + int lp_rtrcredits;
> + /* low water mark */
> + int lp_minrtrcredits;
> + /* alive/dead? */
> + unsigned int lp_alive:1;
> + /* notification outstanding? */
> + unsigned int lp_notify:1;
> + /* outstanding notification for LND? */
> + unsigned int lp_notifylnd:1;
> + /* some thread is handling notification */
> + unsigned int lp_notifying:1;
> + /* SEND event outstanding from ping */
> + unsigned int lp_ping_notsent;
> + /* # times router went dead<->alive */
> + int lp_alive_count;
> + /* bytes queued for sending */
> + long lp_txqnob;
> + /* time of last aliveness news */
> + unsigned long lp_timestamp;
> + /* time of last ping attempt */
> + unsigned long lp_ping_timestamp;
> + /* != 0 if ping reply expected */
> + unsigned long lp_ping_deadline;
> + /* when I was last alive */
> + unsigned long lp_last_alive;
> + /* when lp_ni was queried last time */
> + unsigned long lp_last_query;
> + /* interface peer is on */
> + lnet_ni_t *lp_ni;
> + /* peer's NID */
> + lnet_nid_t lp_nid;
> + /* # refs */
> + int lp_refcount;
> + /* CPT this peer attached on */
> + int lp_cpt;
This new block of declarations is uglier than the original. Don't make
things uglier.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
2015-05-22 13:00 ` Dan Carpenter
@ 2015-05-22 15:12 ` Simmons, James A.
-1 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-22 15:12 UTC (permalink / raw)
To: lustre-devel
>> typedef struct lnet_peer {
>> - struct list_head lp_hashlist; /* chain on peer hash */
>> - struct list_head lp_txq; /* messages blocking for tx credits */
>> - struct list_head lp_rtrq; /* messages blocking for router credits */
>> - struct list_head lp_rtr_list; /* chain on router list */
>> - int lp_txcredits; /* # tx credits available */
>> - int lp_mintxcredits; /* low water mark */
>> - int lp_rtrcredits; /* # router credits */
>> - int lp_minrtrcredits; /* low water mark */
>> - unsigned int lp_alive:1; /* alive/dead? */
>> - unsigned int lp_notify:1; /* notification outstanding? */
>> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
>> - unsigned int lp_notifying:1; /* some thread is handling notification */
>> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
>> - int lp_alive_count; /* # times router went dead<->alive */
>> - long lp_txqnob; /* bytes queued for sending */
>> - unsigned long lp_timestamp; /* time of last aliveness news */
>> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
>> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
>> - unsigned long lp_last_alive; /* when I was last alive */
>> - unsigned long lp_last_query; /* when lp_ni was queried last time */
>> - lnet_ni_t *lp_ni; /* interface peer is on */
>> - lnet_nid_t lp_nid; /* peer's NID */
>> - int lp_refcount; /* # refs */
>> - int lp_cpt; /* CPT this peer attached on */
>> + /* chain on peer hash */
>> + struct list_head lp_hashlist;
>> + /* messages blocking for tx credits */
>> + struct list_head lp_txq;
>> + /* messages blocking for router credits */
>> + struct list_head lp_rtrq;
>> + /* chain on router list */
>> + struct list_head lp_rtr_list;
>> + /* # tx credits available */
>> + int lp_txcredits;
>> + /* low water mark */
>> + int lp_mintxcredits;
>> + /* # router credits */
>> + int lp_rtrcredits;
>> + /* low water mark */
>> + int lp_minrtrcredits;
>> + /* alive/dead? */
>> + unsigned int lp_alive:1;
>> + /* notification outstanding? */
>> + unsigned int lp_notify:1;
>> + /* outstanding notification for LND? */
>> + unsigned int lp_notifylnd:1;
>
>This new block of declarations is uglier than the original. Don't make
>things uglier.
Might be ugly but it makes checkpatch.pl happy. So it is a choice between
making checkpatch.pl happy about staying in the 80 character limit or looking
nice and chekpatch.pl being unhappy.
^ permalink raw reply [flat|nested] 44+ messages in thread* RE: [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
@ 2015-05-22 15:12 ` Simmons, James A.
0 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-22 15:12 UTC (permalink / raw)
To: 'Dan Carpenter', James Simmons
Cc: devel@driverdev.osuosl.org, James Simmons, James Simmons,
Greg Kroah-Hartman, Linux Kernel Mailing List, Oleg Drokin,
HPDD-discuss@ml01.01.org, lustre-devel@lists.lustre.org
>> typedef struct lnet_peer {
>> - struct list_head lp_hashlist; /* chain on peer hash */
>> - struct list_head lp_txq; /* messages blocking for tx credits */
>> - struct list_head lp_rtrq; /* messages blocking for router credits */
>> - struct list_head lp_rtr_list; /* chain on router list */
>> - int lp_txcredits; /* # tx credits available */
>> - int lp_mintxcredits; /* low water mark */
>> - int lp_rtrcredits; /* # router credits */
>> - int lp_minrtrcredits; /* low water mark */
>> - unsigned int lp_alive:1; /* alive/dead? */
>> - unsigned int lp_notify:1; /* notification outstanding? */
>> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
>> - unsigned int lp_notifying:1; /* some thread is handling notification */
>> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
>> - int lp_alive_count; /* # times router went dead<->alive */
>> - long lp_txqnob; /* bytes queued for sending */
>> - unsigned long lp_timestamp; /* time of last aliveness news */
>> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
>> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
>> - unsigned long lp_last_alive; /* when I was last alive */
>> - unsigned long lp_last_query; /* when lp_ni was queried last time */
>> - lnet_ni_t *lp_ni; /* interface peer is on */
>> - lnet_nid_t lp_nid; /* peer's NID */
>> - int lp_refcount; /* # refs */
>> - int lp_cpt; /* CPT this peer attached on */
>> + /* chain on peer hash */
>> + struct list_head lp_hashlist;
>> + /* messages blocking for tx credits */
>> + struct list_head lp_txq;
>> + /* messages blocking for router credits */
>> + struct list_head lp_rtrq;
>> + /* chain on router list */
>> + struct list_head lp_rtr_list;
>> + /* # tx credits available */
>> + int lp_txcredits;
>> + /* low water mark */
>> + int lp_mintxcredits;
>> + /* # router credits */
>> + int lp_rtrcredits;
>> + /* low water mark */
>> + int lp_minrtrcredits;
>> + /* alive/dead? */
>> + unsigned int lp_alive:1;
>> + /* notification outstanding? */
>> + unsigned int lp_notify:1;
>> + /* outstanding notification for LND? */
>> + unsigned int lp_notifylnd:1;
>
>This new block of declarations is uglier than the original. Don't make
>things uglier.
Might be ugly but it makes checkpatch.pl happy. So it is a choice between
making checkpatch.pl happy about staying in the 80 character limit or looking
nice and chekpatch.pl being unhappy.
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
2015-05-22 15:12 ` Simmons, James A.
@ 2015-05-22 15:24 ` Joe Perches
-1 siblings, 0 replies; 44+ messages in thread
From: Joe Perches @ 2015-05-22 15:24 UTC (permalink / raw)
To: lustre-devel
On Fri, 2015-05-22 at 15:12 +0000, Simmons, James A. wrote:
> >> typedef struct lnet_peer {
> >> - struct list_head lp_hashlist; /* chain on peer hash */
> >> - struct list_head lp_txq; /* messages blocking for tx credits */
> >> - struct list_head lp_rtrq; /* messages blocking for router credits */
> >> - struct list_head lp_rtr_list; /* chain on router list */
> >> - int lp_txcredits; /* # tx credits available */
> >> - int lp_mintxcredits; /* low water mark */
> >> - int lp_rtrcredits; /* # router credits */
> >> - int lp_minrtrcredits; /* low water mark */
> >> - unsigned int lp_alive:1; /* alive/dead? */
> >> - unsigned int lp_notify:1; /* notification outstanding? */
> >> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
> >> - unsigned int lp_notifying:1; /* some thread is handling notification */
> >> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
> >> - int lp_alive_count; /* # times router went dead<->alive */
> >> - long lp_txqnob; /* bytes queued for sending */
> >> - unsigned long lp_timestamp; /* time of last aliveness news */
> >> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
> >> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
> >> - unsigned long lp_last_alive; /* when I was last alive */
> >> - unsigned long lp_last_query; /* when lp_ni was queried last time */
> >> - lnet_ni_t *lp_ni; /* interface peer is on */
> >> - lnet_nid_t lp_nid; /* peer's NID */
> >> - int lp_refcount; /* # refs */
> >> - int lp_cpt; /* CPT this peer attached on */
> >> + /* chain on peer hash */
> >> + struct list_head lp_hashlist;
> >> + /* messages blocking for tx credits */
> >> + struct list_head lp_txq;
> >> + /* messages blocking for router credits */
> >> + struct list_head lp_rtrq;
> >> + /* chain on router list */
> >> + struct list_head lp_rtr_list;
> >> + /* # tx credits available */
> >> + int lp_txcredits;
> >> + /* low water mark */
> >> + int lp_mintxcredits;
> >> + /* # router credits */
> >> + int lp_rtrcredits;
> >> + /* low water mark */
> >> + int lp_minrtrcredits;
> >> + /* alive/dead? */
> >> + unsigned int lp_alive:1;
> >> + /* notification outstanding? */
> >> + unsigned int lp_notify:1;
> >> + /* outstanding notification for LND? */
> >> + unsigned int lp_notifylnd:1;
> >
> >This new block of declarations is uglier than the original. Don't make
> >things uglier.
>
> Might be ugly but it makes checkpatch.pl happy. So it is a choice between
> making checkpatch.pl happy about staying in the 80 character limit or looking
> nice and chekpatch.pl being unhappy.
I would choose looking nice every time.
checkpatch is stupid. Please don't let it control you.
Maybe it'd be better to add another --ignore type just for
comments that extend longer than the maximum line length to
checkpatch.
^ permalink raw reply [flat|nested] 44+ messages in thread* Re: [lustre-devel] [PATCH 2/3] staging:lustre: remove kernel defines in userland headers
@ 2015-05-22 15:24 ` Joe Perches
0 siblings, 0 replies; 44+ messages in thread
From: Joe Perches @ 2015-05-22 15:24 UTC (permalink / raw)
To: Simmons, James A.
Cc: 'Dan Carpenter', James Simmons,
devel@driverdev.osuosl.org, James Simmons, James Simmons,
Greg Kroah-Hartman, Linux Kernel Mailing List, Oleg Drokin,
HPDD-discuss@ml01.01.org, lustre-devel@lists.lustre.org
On Fri, 2015-05-22 at 15:12 +0000, Simmons, James A. wrote:
> >> typedef struct lnet_peer {
> >> - struct list_head lp_hashlist; /* chain on peer hash */
> >> - struct list_head lp_txq; /* messages blocking for tx credits */
> >> - struct list_head lp_rtrq; /* messages blocking for router credits */
> >> - struct list_head lp_rtr_list; /* chain on router list */
> >> - int lp_txcredits; /* # tx credits available */
> >> - int lp_mintxcredits; /* low water mark */
> >> - int lp_rtrcredits; /* # router credits */
> >> - int lp_minrtrcredits; /* low water mark */
> >> - unsigned int lp_alive:1; /* alive/dead? */
> >> - unsigned int lp_notify:1; /* notification outstanding? */
> >> - unsigned int lp_notifylnd:1; /* outstanding notification for LND? */
> >> - unsigned int lp_notifying:1; /* some thread is handling notification */
> >> - unsigned int lp_ping_notsent; /* SEND event outstanding from ping */
> >> - int lp_alive_count; /* # times router went dead<->alive */
> >> - long lp_txqnob; /* bytes queued for sending */
> >> - unsigned long lp_timestamp; /* time of last aliveness news */
> >> - unsigned long lp_ping_timestamp; /* time of last ping attempt */
> >> - unsigned long lp_ping_deadline; /* != 0 if ping reply expected */
> >> - unsigned long lp_last_alive; /* when I was last alive */
> >> - unsigned long lp_last_query; /* when lp_ni was queried last time */
> >> - lnet_ni_t *lp_ni; /* interface peer is on */
> >> - lnet_nid_t lp_nid; /* peer's NID */
> >> - int lp_refcount; /* # refs */
> >> - int lp_cpt; /* CPT this peer attached on */
> >> + /* chain on peer hash */
> >> + struct list_head lp_hashlist;
> >> + /* messages blocking for tx credits */
> >> + struct list_head lp_txq;
> >> + /* messages blocking for router credits */
> >> + struct list_head lp_rtrq;
> >> + /* chain on router list */
> >> + struct list_head lp_rtr_list;
> >> + /* # tx credits available */
> >> + int lp_txcredits;
> >> + /* low water mark */
> >> + int lp_mintxcredits;
> >> + /* # router credits */
> >> + int lp_rtrcredits;
> >> + /* low water mark */
> >> + int lp_minrtrcredits;
> >> + /* alive/dead? */
> >> + unsigned int lp_alive:1;
> >> + /* notification outstanding? */
> >> + unsigned int lp_notify:1;
> >> + /* outstanding notification for LND? */
> >> + unsigned int lp_notifylnd:1;
> >
> >This new block of declarations is uglier than the original. Don't make
> >things uglier.
>
> Might be ugly but it makes checkpatch.pl happy. So it is a choice between
> making checkpatch.pl happy about staying in the 80 character limit or looking
> nice and chekpatch.pl being unhappy.
I would choose looking nice every time.
checkpatch is stupid. Please don't let it control you.
Maybe it'd be better to add another --ignore type just for
comments that extend longer than the maximum line length to
checkpatch.
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 3/3] staging:lustre: cleanup libcfs lock handling
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: lustre-devel
Previously with libcfs being built for user land and kernel
space wrappers were created to transparently handle locking.
Now that user land support has been removed we delete all
those locking wrappers with this patch. Many of those changes
landed upstream but some nice cleanups still remain that are
pushed in this patch.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13793
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
---
drivers/staging/lustre/lnet/lnet/acceptor.c | 2 +-
drivers/staging/lustre/lustre/libcfs/fail.c | 2 +-
.../lustre/lustre/libcfs/linux/linux-tracefile.c | 4 +---
drivers/staging/lustre/lustre/libcfs/module.c | 11 ++---------
drivers/staging/lustre/lustre/libcfs/tracefile.c | 2 +-
5 files changed, 6 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index c44d3e0..a7c5785 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -35,9 +35,9 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
+#include <linux/completion.h>
#include "../../include/linux/lnet/lib-lnet.h"
-
static int accept_port = 988;
static int accept_backlog = 127;
static int accept_timeout = 5;
diff --git a/drivers/staging/lustre/lustre/libcfs/fail.c b/drivers/staging/lustre/lustre/libcfs/fail.c
index 92444b0..7b7fc21 100644
--- a/drivers/staging/lustre/lustre/libcfs/fail.c
+++ b/drivers/staging/lustre/lustre/libcfs/fail.c
@@ -41,7 +41,7 @@ EXPORT_SYMBOL(cfs_fail_loc);
unsigned int cfs_fail_val = 0;
EXPORT_SYMBOL(cfs_fail_val);
-wait_queue_head_t cfs_race_waitq;
+DECLARE_WAIT_QUEUE_HEAD(cfs_race_waitq);
EXPORT_SYMBOL(cfs_race_waitq);
int cfs_race_state;
diff --git a/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c b/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
index 483cbc8..eb10e3b 100644
--- a/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
@@ -49,7 +49,7 @@ static unsigned int pages_factor[CFS_TCD_TYPE_MAX] = {
char *cfs_trace_console_buffers[NR_CPUS][CFS_TCD_TYPE_MAX];
-struct rw_semaphore cfs_tracefile_sem;
+static DECLARE_RWSEM(cfs_tracefile_sem);
int cfs_tracefile_init_arch(void)
{
@@ -57,8 +57,6 @@ int cfs_tracefile_init_arch(void)
int j;
struct cfs_trace_cpu_data *tcd;
- init_rwsem(&cfs_tracefile_sem);
-
/* initialize trace_data */
memset(cfs_trace_data, 0, sizeof(cfs_trace_data));
for (i = 0; i < CFS_TCD_TYPE_MAX; i++) {
diff --git a/drivers/staging/lustre/lustre/libcfs/module.c b/drivers/staging/lustre/lustre/libcfs/module.c
index f0ee76a..22fe9cf 100644
--- a/drivers/staging/lustre/lustre/libcfs/module.c
+++ b/drivers/staging/lustre/lustre/libcfs/module.c
@@ -67,8 +67,6 @@ MODULE_DESCRIPTION("Portals v3.1");
MODULE_LICENSE("GPL");
extern struct miscdevice libcfs_dev;
-extern struct rw_semaphore cfs_tracefile_sem;
-extern struct mutex cfs_trace_thread_mutex;
extern struct cfs_wi_sched *cfs_sched_rehash;
extern void libcfs_init_nidstrings(void);
@@ -249,8 +247,8 @@ static int libcfs_psdev_release(unsigned long flags, void *args)
return 0;
}
-static struct rw_semaphore ioctl_list_sem;
-static struct list_head ioctl_list;
+static DECLARE_RWSEM(ioctl_list_sem);
+static LIST_HEAD(ioctl_list);
int libcfs_register_ioctl(struct libcfs_ioctl_handler *hand)
{
@@ -393,11 +391,6 @@ static int init_libcfs_module(void)
libcfs_arch_init();
libcfs_init_nidstrings();
- init_rwsem(&cfs_tracefile_sem);
- mutex_init(&cfs_trace_thread_mutex);
- init_rwsem(&ioctl_list_sem);
- INIT_LIST_HEAD(&ioctl_list);
- init_waitqueue_head(&cfs_race_waitq);
rc = libcfs_debug_init(5 * 1024 * 1024);
if (rc < 0) {
diff --git a/drivers/staging/lustre/lustre/libcfs/tracefile.c b/drivers/staging/lustre/lustre/libcfs/tracefile.c
index c86394f..6ee2adc 100644
--- a/drivers/staging/lustre/lustre/libcfs/tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/tracefile.c
@@ -52,7 +52,7 @@ union cfs_trace_data_union (*cfs_trace_data[TCD_MAX_TYPES])[NR_CPUS] __cacheline
char cfs_tracefile[TRACEFILE_NAME_SIZE];
long long cfs_tracefile_size = CFS_TRACEFILE_SIZE;
static struct tracefiled_ctl trace_tctl;
-struct mutex cfs_trace_thread_mutex;
+static DEFINE_MUTEX(cfs_trace_thread_mutex);
static int thread_running;
static atomic_t cfs_tage_allocated = ATOMIC_INIT(0);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 3/3] staging:lustre: cleanup libcfs lock handling
@ 2015-05-21 22:46 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-21 22:46 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons, James Simmons
Previously with libcfs being built for user land and kernel
space wrappers were created to transparently handle locking.
Now that user land support has been removed we delete all
those locking wrappers with this patch. Many of those changes
landed upstream but some nice cleanups still remain that are
pushed in this patch.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13793
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
---
drivers/staging/lustre/lnet/lnet/acceptor.c | 2 +-
drivers/staging/lustre/lustre/libcfs/fail.c | 2 +-
.../lustre/lustre/libcfs/linux/linux-tracefile.c | 4 +---
drivers/staging/lustre/lustre/libcfs/module.c | 11 ++---------
drivers/staging/lustre/lustre/libcfs/tracefile.c | 2 +-
5 files changed, 6 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index c44d3e0..a7c5785 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -35,9 +35,9 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
+#include <linux/completion.h>
#include "../../include/linux/lnet/lib-lnet.h"
-
static int accept_port = 988;
static int accept_backlog = 127;
static int accept_timeout = 5;
diff --git a/drivers/staging/lustre/lustre/libcfs/fail.c b/drivers/staging/lustre/lustre/libcfs/fail.c
index 92444b0..7b7fc21 100644
--- a/drivers/staging/lustre/lustre/libcfs/fail.c
+++ b/drivers/staging/lustre/lustre/libcfs/fail.c
@@ -41,7 +41,7 @@ EXPORT_SYMBOL(cfs_fail_loc);
unsigned int cfs_fail_val = 0;
EXPORT_SYMBOL(cfs_fail_val);
-wait_queue_head_t cfs_race_waitq;
+DECLARE_WAIT_QUEUE_HEAD(cfs_race_waitq);
EXPORT_SYMBOL(cfs_race_waitq);
int cfs_race_state;
diff --git a/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c b/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
index 483cbc8..eb10e3b 100644
--- a/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/linux/linux-tracefile.c
@@ -49,7 +49,7 @@ static unsigned int pages_factor[CFS_TCD_TYPE_MAX] = {
char *cfs_trace_console_buffers[NR_CPUS][CFS_TCD_TYPE_MAX];
-struct rw_semaphore cfs_tracefile_sem;
+static DECLARE_RWSEM(cfs_tracefile_sem);
int cfs_tracefile_init_arch(void)
{
@@ -57,8 +57,6 @@ int cfs_tracefile_init_arch(void)
int j;
struct cfs_trace_cpu_data *tcd;
- init_rwsem(&cfs_tracefile_sem);
-
/* initialize trace_data */
memset(cfs_trace_data, 0, sizeof(cfs_trace_data));
for (i = 0; i < CFS_TCD_TYPE_MAX; i++) {
diff --git a/drivers/staging/lustre/lustre/libcfs/module.c b/drivers/staging/lustre/lustre/libcfs/module.c
index f0ee76a..22fe9cf 100644
--- a/drivers/staging/lustre/lustre/libcfs/module.c
+++ b/drivers/staging/lustre/lustre/libcfs/module.c
@@ -67,8 +67,6 @@ MODULE_DESCRIPTION("Portals v3.1");
MODULE_LICENSE("GPL");
extern struct miscdevice libcfs_dev;
-extern struct rw_semaphore cfs_tracefile_sem;
-extern struct mutex cfs_trace_thread_mutex;
extern struct cfs_wi_sched *cfs_sched_rehash;
extern void libcfs_init_nidstrings(void);
@@ -249,8 +247,8 @@ static int libcfs_psdev_release(unsigned long flags, void *args)
return 0;
}
-static struct rw_semaphore ioctl_list_sem;
-static struct list_head ioctl_list;
+static DECLARE_RWSEM(ioctl_list_sem);
+static LIST_HEAD(ioctl_list);
int libcfs_register_ioctl(struct libcfs_ioctl_handler *hand)
{
@@ -393,11 +391,6 @@ static int init_libcfs_module(void)
libcfs_arch_init();
libcfs_init_nidstrings();
- init_rwsem(&cfs_tracefile_sem);
- mutex_init(&cfs_trace_thread_mutex);
- init_rwsem(&ioctl_list_sem);
- INIT_LIST_HEAD(&ioctl_list);
- init_waitqueue_head(&cfs_race_waitq);
rc = libcfs_debug_init(5 * 1024 * 1024);
if (rc < 0) {
diff --git a/drivers/staging/lustre/lustre/libcfs/tracefile.c b/drivers/staging/lustre/lustre/libcfs/tracefile.c
index c86394f..6ee2adc 100644
--- a/drivers/staging/lustre/lustre/libcfs/tracefile.c
+++ b/drivers/staging/lustre/lustre/libcfs/tracefile.c
@@ -52,7 +52,7 @@ union cfs_trace_data_union (*cfs_trace_data[TCD_MAX_TYPES])[NR_CPUS] __cacheline
char cfs_tracefile[TRACEFILE_NAME_SIZE];
long long cfs_tracefile_size = CFS_TRACEFILE_SIZE;
static struct tracefiled_ctl trace_tctl;
-struct mutex cfs_trace_thread_mutex;
+static DEFINE_MUTEX(cfs_trace_thread_mutex);
static int thread_running;
static atomic_t cfs_tage_allocated = ATOMIC_INIT(0);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 0/6] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
Since libcfs no longer builds for user land we can
move the TCPIP abstraction that exist to the LNET
layer which is the only place that uses it. Also
the migrated code will use native linux kernel
APIs directly instead of with wrappers.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13760
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
James Simmons (6):
staging:lustre:remove useless libcfs_sock_release
staging:lustre:remove useless libcfs_sock_abort_accept
staging:lustre: rename tcpip handling functions to lnet_* prefix
staging:lustre: use available kernel wrappers in lib-socket.c
staging:lustre: style cleanups for lib-socket.c
staging:lustre: Update license and copyright for lib-socket.c
.../staging/lustre/include/linux/libcfs/libcfs.h | 17 -
.../staging/lustre/include/linux/lnet/lib-lnet.h | 16 +
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 17 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 6 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 15 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 43 ++--
drivers/staging/lustre/lnet/lnet/acceptor.c | 55 ++--
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 323 +++++++++-----------
11 files changed, 236 insertions(+), 281 deletions(-)
^ permalink raw reply [flat|nested] 44+ messages in thread
* [PATCH 0/6] staging:lustre: remove tcpip abstraction from libcfs
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons, James Simmons
Since libcfs no longer builds for user land we can
move the TCPIP abstraction that exist to the LNET
layer which is the only place that uses it. Also
the migrated code will use native linux kernel
APIs directly instead of with wrappers.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/13760
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Isaac Huang <he.huang@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
James Simmons (6):
staging:lustre:remove useless libcfs_sock_release
staging:lustre:remove useless libcfs_sock_abort_accept
staging:lustre: rename tcpip handling functions to lnet_* prefix
staging:lustre: use available kernel wrappers in lib-socket.c
staging:lustre: style cleanups for lib-socket.c
staging:lustre: Update license and copyright for lib-socket.c
.../staging/lustre/include/linux/libcfs/libcfs.h | 17 -
.../staging/lustre/include/linux/lnet/lib-lnet.h | 16 +
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 17 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 6 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 15 +-
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 43 ++--
drivers/staging/lustre/lnet/lnet/acceptor.c | 55 ++--
drivers/staging/lustre/lnet/lnet/config.c | 21 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 323 +++++++++-----------
11 files changed, 236 insertions(+), 281 deletions(-)
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 1/6] staging:lustre:remove useless libcfs_sock_release
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
There is no reason to have a one line exported function
libcfs_sock_release. Instead we can call sock_release directly.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 1 -
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 8 ++++----
drivers/staging/lustre/lnet/lnet/lib-socket.c | 8 --------
5 files changed, 6 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 947df7e..7a7a525 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -103,7 +103,6 @@ int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-void libcfs_sock_release(struct socket *sock);
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index 7586b7e..e642752 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -1386,7 +1386,7 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
LIBCFS_FREE(conn, sizeof(*conn));
failed_0:
- libcfs_sock_release(sock);
+ sock_release(sock);
return rc;
}
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index c54c995..f3b4923 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -453,7 +453,7 @@ ksocknal_connsock_decref(ksock_conn_t *conn)
LASSERT(atomic_read(&conn->ksnc_sock_refcount) > 0);
if (atomic_dec_and_test(&conn->ksnc_sock_refcount)) {
LASSERT(conn->ksnc_closing);
- libcfs_sock_release(conn->ksnc_sock);
+ sock_release(conn->ksnc_sock);
conn->ksnc_sock = NULL;
ksocknal_finalize_zcreq(conn);
}
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 72fd1bf..14673ef 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -197,7 +197,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
goto failed;
failed_sock:
- libcfs_sock_release(sock);
+ sock_release(sock);
failed:
lnet_connect_console_error(rc, peer_nid, peer_ip, peer_port);
return rc;
@@ -379,7 +379,7 @@ lnet_acceptor(void *arg)
/* maybe we're waken up with libcfs_sock_abort_accept() */
if (lnet_acceptor_state.pta_shutdown) {
- libcfs_sock_release(newsock);
+ sock_release(newsock);
break;
}
@@ -410,10 +410,10 @@ lnet_acceptor(void *arg)
continue;
failed:
- libcfs_sock_release(newsock);
+ sock_release(newsock);
}
- libcfs_sock_release(lnet_acceptor_state.pta_sock);
+ sock_release(lnet_acceptor_state.pta_sock);
lnet_acceptor_state.pta_sock = NULL;
CDEBUG(D_NET, "Acceptor stopping\n");
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 7f80612..259db61 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -614,11 +614,3 @@ libcfs_sock_connect (struct socket **sockp, int *fatal,
}
EXPORT_SYMBOL(libcfs_sock_connect);
-
-void
-libcfs_sock_release (struct socket *sock)
-{
- sock_release(sock);
-}
-
-EXPORT_SYMBOL(libcfs_sock_release);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 1/6] staging:lustre:remove useless libcfs_sock_release
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
There is no reason to have a one line exported function
libcfs_sock_release. Instead we can call sock_release directly.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 1 -
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.h | 2 +-
drivers/staging/lustre/lnet/lnet/acceptor.c | 8 ++++----
drivers/staging/lustre/lnet/lnet/lib-socket.c | 8 --------
5 files changed, 6 insertions(+), 15 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 947df7e..7a7a525 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -103,7 +103,6 @@ int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-void libcfs_sock_release(struct socket *sock);
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index 7586b7e..e642752 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -1386,7 +1386,7 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
LIBCFS_FREE(conn, sizeof(*conn));
failed_0:
- libcfs_sock_release(sock);
+ sock_release(sock);
return rc;
}
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
index c54c995..f3b4923 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.h
@@ -453,7 +453,7 @@ ksocknal_connsock_decref(ksock_conn_t *conn)
LASSERT(atomic_read(&conn->ksnc_sock_refcount) > 0);
if (atomic_dec_and_test(&conn->ksnc_sock_refcount)) {
LASSERT(conn->ksnc_closing);
- libcfs_sock_release(conn->ksnc_sock);
+ sock_release(conn->ksnc_sock);
conn->ksnc_sock = NULL;
ksocknal_finalize_zcreq(conn);
}
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 72fd1bf..14673ef 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -197,7 +197,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
goto failed;
failed_sock:
- libcfs_sock_release(sock);
+ sock_release(sock);
failed:
lnet_connect_console_error(rc, peer_nid, peer_ip, peer_port);
return rc;
@@ -379,7 +379,7 @@ lnet_acceptor(void *arg)
/* maybe we're waken up with libcfs_sock_abort_accept() */
if (lnet_acceptor_state.pta_shutdown) {
- libcfs_sock_release(newsock);
+ sock_release(newsock);
break;
}
@@ -410,10 +410,10 @@ lnet_acceptor(void *arg)
continue;
failed:
- libcfs_sock_release(newsock);
+ sock_release(newsock);
}
- libcfs_sock_release(lnet_acceptor_state.pta_sock);
+ sock_release(lnet_acceptor_state.pta_sock);
lnet_acceptor_state.pta_sock = NULL;
CDEBUG(D_NET, "Acceptor stopping\n");
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 7f80612..259db61 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -614,11 +614,3 @@ libcfs_sock_connect (struct socket **sockp, int *fatal,
}
EXPORT_SYMBOL(libcfs_sock_connect);
-
-void
-libcfs_sock_release (struct socket *sock)
-{
- sock_release(sock);
-}
-
-EXPORT_SYMBOL(libcfs_sock_release);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 2/6] staging:lustre:remove useless libcfs_sock_abort_accept
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
Another one of those silly one line wrappers which is not
needed. Replace libcfs_sock_abort_accept wrapper with a
direct call to wake_up_all on the lnet_acceptor_state sock.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 1 -
drivers/staging/lustre/lnet/lnet/acceptor.c | 4 ++--
drivers/staging/lustre/lnet/lnet/lib-socket.c | 8 --------
3 files changed, 2 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 7a7a525..854ba6a 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -94,7 +94,6 @@ int libcfs_ipif_enumerate(char ***names);
void libcfs_ipif_free_enumeration(char **names, int n);
int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-void libcfs_sock_abort_accept(struct socket *sock);
int libcfs_sock_connect(struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
__u32 peer_ip, int peer_port);
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 14673ef..8c6726a 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -377,7 +377,7 @@ lnet_acceptor(void *arg)
continue;
}
- /* maybe we're waken up with libcfs_sock_abort_accept() */
+ /* maybe the LNet acceptor thread has been waken */
if (lnet_acceptor_state.pta_shutdown) {
sock_release(newsock);
break;
@@ -493,7 +493,7 @@ lnet_acceptor_stop(void)
return;
lnet_acceptor_state.pta_shutdown = 1;
- libcfs_sock_abort_accept(lnet_acceptor_state.pta_sock);
+ wake_up_all(sk_sleep(lnet_acceptor_state.pta_sock->sk));
/* block until acceptor signals exit */
wait_for_completion(&lnet_acceptor_state.pta_signal);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 259db61..bb8d9c2 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -568,14 +568,6 @@ libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
EXPORT_SYMBOL(libcfs_sock_accept);
-void
-libcfs_sock_abort_accept (struct socket *sock)
-{
- wake_up_all(sk_sleep(sock->sk));
-}
-
-EXPORT_SYMBOL(libcfs_sock_abort_accept);
-
int
libcfs_sock_connect (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 2/6] staging:lustre:remove useless libcfs_sock_abort_accept
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
Another one of those silly one line wrappers which is not
needed. Replace libcfs_sock_abort_accept wrapper with a
direct call to wake_up_all on the lnet_acceptor_state sock.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 1 -
drivers/staging/lustre/lnet/lnet/acceptor.c | 4 ++--
drivers/staging/lustre/lnet/lnet/lib-socket.c | 8 --------
3 files changed, 2 insertions(+), 11 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 7a7a525..854ba6a 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -94,7 +94,6 @@ int libcfs_ipif_enumerate(char ***names);
void libcfs_ipif_free_enumeration(char **names, int n);
int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-void libcfs_sock_abort_accept(struct socket *sock);
int libcfs_sock_connect(struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
__u32 peer_ip, int peer_port);
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 14673ef..8c6726a 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -377,7 +377,7 @@ lnet_acceptor(void *arg)
continue;
}
- /* maybe we're waken up with libcfs_sock_abort_accept() */
+ /* maybe the LNet acceptor thread has been waken */
if (lnet_acceptor_state.pta_shutdown) {
sock_release(newsock);
break;
@@ -493,7 +493,7 @@ lnet_acceptor_stop(void)
return;
lnet_acceptor_state.pta_shutdown = 1;
- libcfs_sock_abort_accept(lnet_acceptor_state.pta_sock);
+ wake_up_all(sk_sleep(lnet_acceptor_state.pta_sock->sk));
/* block until acceptor signals exit */
wait_for_completion(&lnet_acceptor_state.pta_signal);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 259db61..bb8d9c2 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -568,14 +568,6 @@ libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
EXPORT_SYMBOL(libcfs_sock_accept);
-void
-libcfs_sock_abort_accept (struct socket *sock)
-{
- wake_up_all(sk_sleep(sock->sk));
-}
-
-EXPORT_SYMBOL(libcfs_sock_abort_accept);
-
int
libcfs_sock_connect (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 3/6] staging:lustre: rename tcpip handling functions to lnet_* prefix
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
With all the TCPIP handling done in the lnet layer we should
rename all the functions with the prefix lnet_*. One other
change done was changing the remove argument of lnet_sock_getaddr
from a int to a bool.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 15 -----
.../staging/lustre/include/linux/lnet/lib-lnet.h | 16 +++++
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 11 ++--
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 6 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 15 ++---
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 43 ++++++-------
drivers/staging/lustre/lnet/lnet/acceptor.c | 43 ++++++-------
drivers/staging/lustre/lnet/lnet/config.c | 21 +++----
drivers/staging/lustre/lnet/lnet/lib-socket.c | 66 +++++++++----------
10 files changed, 110 insertions(+), 128 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 854ba6a..f54bab6 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -88,21 +88,6 @@ static inline int __is_po2(unsigned long long val)
int libcfs_arch_init(void);
void libcfs_arch_cleanup(void);
-/* libcfs tcpip */
-int libcfs_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
-int libcfs_ipif_enumerate(char ***names);
-void libcfs_ipif_free_enumeration(char **names, int n);
-int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
-int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-int libcfs_sock_connect(struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port);
-int libcfs_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
-int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
-int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
-int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
-int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
#define LNET_ACCEPTOR_MAX_RESERVED_PORT 1023
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index 0038d29..5175588 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -846,6 +846,22 @@ int lnet_acceptor_port(void);
int lnet_acceptor_start(void);
void lnet_acceptor_stop(void);
+int lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
+int lnet_ipif_enumerate(char ***names);
+void lnet_ipif_free_enumeration(char **names, int n);
+int lnet_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
+int lnet_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
+int lnet_sock_getaddr(struct socket *socket, bool remote, __u32 *ip, int *port);
+int lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
+int lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
+
+int lnet_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
+int lnet_sock_accept(struct socket **newsockp, struct socket *sock);
+int lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port);
+void libcfs_sock_release(struct socket *sock);
+
void lnet_get_tunables(void);
int lnet_peers_start_down(void);
int lnet_peer_buffer_credits(lnet_ni_t *ni);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index 3bad441..be4de56 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -2620,7 +2620,7 @@ static kib_dev_t *kiblnd_create_dev(char *ifname)
int up;
int rc;
- rc = libcfs_ipif_query(ifname, &up, &ip, &netmask);
+ rc = lnet_ipif_query(ifname, &up, &ip, &netmask);
if (rc != 0) {
CERROR("Can't query IPoIB interface %s: %d\n",
ifname, rc);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index e642752..a27dddf 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -968,7 +968,7 @@ ksocknal_accept(lnet_ni_t *ni, struct socket *sock)
__u32 peer_ip;
int peer_port;
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
LIBCFS_ALLOC(cr, sizeof(*cr));
@@ -2594,7 +2594,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
int rc;
int n;
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Can't enumerate interfaces: %d\n", n);
return n;
@@ -2608,7 +2608,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &mask);
+ rc = lnet_ipif_query(names[i], &up, &ip, &mask);
if (rc != 0) {
CWARN("Can't get interface %s info: %d\n",
names[i], rc);
@@ -2634,7 +2634,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
j++;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
if (j == 0)
CERROR("Can't find any usable interfaces\n");
@@ -2796,8 +2796,7 @@ ksocknal_startup(lnet_ni_t *ni)
if (ni->ni_interfaces[i] == NULL)
break;
- rc = libcfs_ipif_query(
- ni->ni_interfaces[i], &up,
+ rc = lnet_ipif_query(ni->ni_interfaces[i], &up,
&net->ksnn_interfaces[i].ksni_ipaddr,
&net->ksnn_interfaces[i].ksni_netmask);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
index fa7ad88..352f03e 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
@@ -1707,7 +1707,7 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
timeout = active ? *ksocknal_tunables.ksnd_timeout :
lnet_acceptor_timeout();
- rc = libcfs_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -1726,8 +1726,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &hello->kshm_version,
- sizeof(hello->kshm_version), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_version,
+ sizeof(hello->kshm_version), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index f5e8ab0..589f0d6 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -39,9 +39,8 @@
int
ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
{
- int rc = libcfs_sock_getaddr(conn->ksnc_sock, 1,
- &conn->ksnc_ipaddr,
- &conn->ksnc_port);
+ int rc = lnet_sock_getaddr(conn->ksnc_sock, 1, &conn->ksnc_ipaddr,
+ &conn->ksnc_port);
/* Didn't need the {get,put}connsock dance to deref ksnc_sock... */
LASSERT(!conn->ksnc_closing);
@@ -51,8 +50,7 @@ ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
return rc;
}
- rc = libcfs_sock_getaddr(conn->ksnc_sock, 0,
- &conn->ksnc_myipaddr, NULL);
+ rc = lnet_sock_getaddr(conn->ksnc_sock, 0, &conn->ksnc_myipaddr, NULL);
if (rc != 0) {
CERROR("Error %d getting sock local IP\n", rc);
return rc;
@@ -436,7 +434,7 @@ ksocknal_lib_get_conn_tunables(ksock_conn_t *conn, int *txmem, int *rxmem, int *
return -ESHUTDOWN;
}
- rc = libcfs_sock_getbuf(sock, txmem, rxmem);
+ rc = lnet_sock_getbuf(sock, txmem, rxmem);
if (rc == 0) {
len = sizeof(*nagle);
rc = kernel_getsockopt(sock, SOL_TCP, TCP_NODELAY,
@@ -498,9 +496,8 @@ ksocknal_lib_setup_sock(struct socket *sock)
}
}
- rc = libcfs_sock_setbuf(sock,
- *ksocknal_tunables.ksnd_tx_buffer_size,
- *ksocknal_tunables.ksnd_rx_buffer_size);
+ rc = lnet_sock_setbuf(sock, *ksocknal_tunables.ksnd_tx_buffer_size,
+ *ksocknal_tunables.ksnd_rx_buffer_size);
if (rc != 0) {
CERROR("Can't set buffer tx %d, rx %d buffers: %d\n",
*ksocknal_tunables.ksnd_tx_buffer_size,
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
index 8596581..48197eb 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
@@ -495,9 +495,7 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hdr->msg.hello.type = cpu_to_le32 (hello->kshm_ctype);
hdr->msg.hello.incarnation = cpu_to_le64 (hello->kshm_src_incarnation);
- rc = libcfs_sock_write(sock, hdr, sizeof(*hdr),
- lnet_acceptor_timeout());
-
+ rc = lnet_sock_write(sock, hdr, sizeof(*hdr), lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
rc, &conn->ksnc_ipaddr, conn->ksnc_port);
@@ -511,9 +509,9 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hello->kshm_ips[i] = __cpu_to_le32 (hello->kshm_ips[i]);
}
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -544,9 +542,8 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
LNET_UNLOCK();
}
- rc = libcfs_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
- lnet_acceptor_timeout());
-
+ rc = lnet_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
rc, &conn->ksnc_ipaddr, conn->ksnc_port);
@@ -556,9 +553,9 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -583,9 +580,9 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
return -ENOMEM;
}
- rc = libcfs_sock_read(sock, &hdr->src_nid,
- sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
- timeout);
+ rc = lnet_sock_read(sock, &hdr->src_nid,
+ sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
+ timeout);
if (rc != 0) {
CERROR("Error %d reading rest of HELLO hdr from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -619,8 +616,8 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
if (hello->kshm_nips == 0)
goto out;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -656,10 +653,10 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
else
conn->ksnc_flip = 1;
- rc = libcfs_sock_read(sock, &hello->kshm_src_nid,
- offsetof(ksock_hello_msg_t, kshm_ips) -
- offsetof(ksock_hello_msg_t, kshm_src_nid),
- timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_src_nid,
+ offsetof(ksock_hello_msg_t, kshm_ips) -
+ offsetof(ksock_hello_msg_t, kshm_src_nid),
+ timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -687,8 +684,8 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 8c6726a..d323899 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -155,9 +155,8 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
--port) {
/* Iterate through reserved ports. */
- rc = libcfs_sock_connect(&sock, &fatal,
- local_ip, port,
- peer_ip, peer_port);
+ rc = lnet_sock_connect(&sock, &fatal, local_ip, port, peer_ip,
+ peer_port);
if (rc != 0) {
if (fatal)
goto failed;
@@ -184,8 +183,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
lnet_net_unlock(LNET_LOCK_EX);
}
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
goto failed_sock;
@@ -220,7 +218,7 @@ lnet_accept(struct socket *sock, __u32 magic)
LASSERT(sizeof(cr) <= 16); /* not too big for the stack */
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
if (!lnet_accept_magic(magic, LNET_PROTO_ACCEPTOR_MAGIC)) {
@@ -234,8 +232,8 @@ lnet_accept(struct socket *sock, __u32 magic)
memset(&cr, 0, sizeof(cr));
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
+ accept_timeout);
if (rc != 0)
CERROR("Error sending magic+version in response to LNET magic from %pI4h: %d\n",
@@ -257,9 +255,8 @@ lnet_accept(struct socket *sock, __u32 magic)
flip = (magic != LNET_PROTO_ACCEPTOR_MAGIC);
- rc = libcfs_sock_read(sock, &cr.acr_version,
- sizeof(cr.acr_version),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_version, sizeof(cr.acr_version),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request version from %pI4h\n",
rc, &peer_ip);
@@ -280,19 +277,17 @@ lnet_accept(struct socket *sock, __u32 magic)
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
-
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
CERROR("Error sending magic+version in response to version %d from %pI4h: %d\n",
peer_version, &peer_ip, rc);
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &cr.acr_nid,
- sizeof(cr) -
- offsetof(lnet_acceptor_connreq_t, acr_nid),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_nid,
+ sizeof(cr) -
+ offsetof(lnet_acceptor_connreq_t, acr_nid),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
rc, &peer_ip);
@@ -343,8 +338,8 @@ lnet_acceptor(void *arg)
cfs_block_allsigs();
- rc = libcfs_sock_listen(&lnet_acceptor_state.pta_sock,
- 0, accept_port, accept_backlog);
+ rc = lnet_sock_listen(&lnet_acceptor_state.pta_sock, 0, accept_port,
+ accept_backlog);
if (rc != 0) {
if (rc == -EADDRINUSE)
LCONSOLE_ERROR_MSG(0x122, "Can't start acceptor on port %d: port already in use\n",
@@ -367,7 +362,7 @@ lnet_acceptor(void *arg)
while (!lnet_acceptor_state.pta_shutdown) {
- rc = libcfs_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
+ rc = lnet_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
if (rc != 0) {
if (rc != -EAGAIN) {
CWARN("Accept error %d: pausing...\n", rc);
@@ -383,7 +378,7 @@ lnet_acceptor(void *arg)
break;
}
- rc = libcfs_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
if (rc != 0) {
CERROR("Can't determine new connection's address\n");
goto failed;
@@ -395,8 +390,8 @@ lnet_acceptor(void *arg)
goto failed;
}
- rc = libcfs_sock_read(newsock, &magic, sizeof(magic),
- accept_timeout);
+ rc = lnet_sock_read(newsock, &magic, sizeof(magic),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
rc, &peer_ip);
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index 2dc4c4a..d8c8f2e 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -1118,7 +1118,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
__u32 *ipaddrs2;
int nip;
char **ifnames;
- int nif = libcfs_ipif_enumerate(&ifnames);
+ int nif = lnet_ipif_enumerate(&ifnames);
int i;
int rc;
@@ -1128,7 +1128,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
LIBCFS_ALLOC(ipaddrs, nif * sizeof(*ipaddrs));
if (ipaddrs == NULL) {
CERROR("Can't allocate ipaddrs[%d]\n", nif);
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
return -ENOMEM;
}
@@ -1136,8 +1136,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
if (!strcmp(ifnames[i], "lo"))
continue;
- rc = libcfs_ipif_query(ifnames[i], &up,
- &ipaddrs[nip], &netmask);
+ rc = lnet_ipif_query(ifnames[i], &up, &ipaddrs[nip], &netmask);
if (rc != 0) {
CWARN("Can't query interface %s: %d\n",
ifnames[i], rc);
@@ -1153,7 +1152,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
nip++;
}
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
if (nip == nif) {
*ipaddrsp = ipaddrs;
@@ -1237,8 +1236,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return -EPERM;
}
- rc = libcfs_ipif_query(ni->ni_interfaces[0],
- &up, &ip, &netmask);
+ rc = lnet_ipif_query(ni->ni_interfaces[0], &up, &ip, &netmask);
if (rc != 0) {
CERROR("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), ni->ni_interfaces[0], rc);
@@ -1255,7 +1253,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return 0;
}
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Net %s can't enumerate interfaces: %d\n",
libcfs_net2str(net), n);
@@ -1266,8 +1264,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &netmask);
-
+ rc = lnet_ipif_query(names[i], &up, &ip, &netmask);
if (rc != 0) {
CWARN("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), names[i], rc);
@@ -1280,13 +1277,13 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
continue;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
ni->ni_nid = LNET_MKNID(net, ip);
return 0;
}
CERROR("Net %s can't find any interfaces\n", libcfs_net2str(net));
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
return -ENOENT;
}
EXPORT_SYMBOL(lnet_set_ip_niaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index bb8d9c2..2e87168 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -45,7 +45,7 @@
#include <linux/syscalls.h>
static int
-libcfs_sock_ioctl(int cmd, unsigned long arg)
+lnet_sock_ioctl(int cmd, unsigned long arg)
{
mm_segment_t oldmm = get_fs();
struct socket *sock;
@@ -76,7 +76,7 @@ out:
}
int
-libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
+lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
{
struct ifreq ifr;
int nob;
@@ -92,8 +92,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
strcpy(ifr.ifr_name, name);
- rc = libcfs_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get flags for interface %s\n", name);
return rc;
@@ -110,8 +109,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
strcpy(ifr.ifr_name, name);
ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get IP address for interface %s\n", name);
return rc;
@@ -122,8 +120,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
strcpy(ifr.ifr_name, name);
ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get netmask for interface %s\n", name);
return rc;
@@ -135,10 +132,10 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return 0;
}
-EXPORT_SYMBOL(libcfs_ipif_query);
+EXPORT_SYMBOL(lnet_ipif_query);
int
-libcfs_ipif_enumerate (char ***namesp)
+lnet_ipif_enumerate (char ***namesp)
{
/* Allocate and fill in 'names', returning # interfaces/error */
char **names;
@@ -172,8 +169,7 @@ libcfs_ipif_enumerate (char ***namesp)
ifc.ifc_buf = (char *)ifr;
ifc.ifc_len = nalloc * sizeof(*ifr);
- rc = libcfs_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
-
+ rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
if (rc < 0) {
CERROR ("Error %d enumerating interfaces\n", rc);
goto out1;
@@ -226,17 +222,17 @@ libcfs_ipif_enumerate (char ***namesp)
out2:
if (rc < 0)
- libcfs_ipif_free_enumeration(names, nfound);
+ lnet_ipif_free_enumeration(names, nfound);
out1:
LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
out0:
return rc;
}
-EXPORT_SYMBOL(libcfs_ipif_enumerate);
+EXPORT_SYMBOL(lnet_ipif_enumerate);
void
-libcfs_ipif_free_enumeration (char **names, int n)
+lnet_ipif_free_enumeration (char **names, int n)
{
int i;
@@ -248,10 +244,10 @@ libcfs_ipif_free_enumeration (char **names, int n)
LIBCFS_FREE(names, n * sizeof(*names));
}
-EXPORT_SYMBOL(libcfs_ipif_free_enumeration);
+EXPORT_SYMBOL(lnet_ipif_free_enumeration);
int
-libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
{
int rc;
long ticks = timeout * HZ;
@@ -310,10 +306,10 @@ libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_write);
+EXPORT_SYMBOL(lnet_sock_write);
int
-libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
{
int rc;
long ticks = timeout * HZ;
@@ -366,10 +362,10 @@ libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
}
}
-EXPORT_SYMBOL(libcfs_sock_read);
+EXPORT_SYMBOL(lnet_sock_read);
static int
-libcfs_sock_create (struct socket **sockp, int *fatal,
+lnet_sock_create (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port)
{
struct sockaddr_in locaddr;
@@ -424,7 +420,7 @@ libcfs_sock_create (struct socket **sockp, int *fatal,
}
int
-libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
+lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
{
int option;
int rc;
@@ -454,10 +450,10 @@ libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_setbuf);
+EXPORT_SYMBOL(lnet_sock_setbuf);
int
-libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
+lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
{
struct sockaddr_in sin;
int len = sizeof (sin);
@@ -480,10 +476,10 @@ libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_getaddr);
+EXPORT_SYMBOL(lnet_sock_getaddr);
int
-libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
+lnet_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
{
if (txbufsize != NULL) {
@@ -497,16 +493,16 @@ libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_getbuf);
+EXPORT_SYMBOL(lnet_sock_getbuf);
int
-libcfs_sock_listen (struct socket **sockp,
+lnet_sock_listen (struct socket **sockp,
__u32 local_ip, int local_port, int backlog)
{
int fatal;
int rc;
- rc = libcfs_sock_create(sockp, &fatal, local_ip, local_port);
+ rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
if (rc != 0) {
if (!fatal)
CERROR("Can't create socket: port %d already in use\n",
@@ -523,10 +519,10 @@ libcfs_sock_listen (struct socket **sockp,
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_listen);
+EXPORT_SYMBOL(lnet_sock_listen);
int
-libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
+lnet_sock_accept (struct socket **newsockp, struct socket *sock)
{
wait_queue_t wait;
struct socket *newsock;
@@ -566,17 +562,17 @@ libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_accept);
+EXPORT_SYMBOL(lnet_sock_accept);
int
-libcfs_sock_connect (struct socket **sockp, int *fatal,
+lnet_sock_connect (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
__u32 peer_ip, int peer_port)
{
struct sockaddr_in srvaddr;
int rc;
- rc = libcfs_sock_create(sockp, fatal, local_ip, local_port);
+ rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
if (rc != 0)
return rc;
@@ -605,4 +601,4 @@ libcfs_sock_connect (struct socket **sockp, int *fatal,
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_connect);
+EXPORT_SYMBOL(lnet_sock_connect);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 3/6] staging:lustre: rename tcpip handling functions to lnet_* prefix
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
With all the TCPIP handling done in the lnet layer we should
rename all the functions with the prefix lnet_*. One other
change done was changing the remove argument of lnet_sock_getaddr
from a int to a bool.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/include/linux/libcfs/libcfs.h | 15 -----
.../staging/lustre/include/linux/lnet/lib-lnet.h | 16 +++++
.../staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c | 2 +-
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 11 ++--
.../staging/lustre/lnet/klnds/socklnd/socklnd_cb.c | 6 +-
.../lustre/lnet/klnds/socklnd/socklnd_lib-linux.c | 15 ++---
.../lustre/lnet/klnds/socklnd/socklnd_proto.c | 43 ++++++-------
drivers/staging/lustre/lnet/lnet/acceptor.c | 43 ++++++-------
drivers/staging/lustre/lnet/lnet/config.c | 21 +++----
drivers/staging/lustre/lnet/lnet/lib-socket.c | 66 +++++++++----------
10 files changed, 110 insertions(+), 128 deletions(-)
diff --git a/drivers/staging/lustre/include/linux/libcfs/libcfs.h b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
index 854ba6a..f54bab6 100644
--- a/drivers/staging/lustre/include/linux/libcfs/libcfs.h
+++ b/drivers/staging/lustre/include/linux/libcfs/libcfs.h
@@ -88,21 +88,6 @@ static inline int __is_po2(unsigned long long val)
int libcfs_arch_init(void);
void libcfs_arch_cleanup(void);
-/* libcfs tcpip */
-int libcfs_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
-int libcfs_ipif_enumerate(char ***names);
-void libcfs_ipif_free_enumeration(char **names, int n);
-int libcfs_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
-int libcfs_sock_accept(struct socket **newsockp, struct socket *sock);
-int libcfs_sock_connect(struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port);
-int libcfs_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
-int libcfs_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
-int libcfs_sock_getaddr(struct socket *socket, int remote, __u32 *ip, int *port);
-int libcfs_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
-int libcfs_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
-
/* need both kernel and user-land acceptor */
#define LNET_ACCEPTOR_MIN_RESERVED_PORT 512
#define LNET_ACCEPTOR_MAX_RESERVED_PORT 1023
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
index 0038d29..5175588 100644
--- a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -846,6 +846,22 @@ int lnet_acceptor_port(void);
int lnet_acceptor_start(void);
void lnet_acceptor_stop(void);
+int lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask);
+int lnet_ipif_enumerate(char ***names);
+void lnet_ipif_free_enumeration(char **names, int n);
+int lnet_sock_setbuf(struct socket *socket, int txbufsize, int rxbufsize);
+int lnet_sock_getbuf(struct socket *socket, int *txbufsize, int *rxbufsize);
+int lnet_sock_getaddr(struct socket *socket, bool remote, __u32 *ip, int *port);
+int lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout);
+int lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout);
+
+int lnet_sock_listen(struct socket **sockp, __u32 ip, int port, int backlog);
+int lnet_sock_accept(struct socket **newsockp, struct socket *sock);
+int lnet_sock_connect(struct socket **sockp, int *fatal,
+ __u32 local_ip, int local_port,
+ __u32 peer_ip, int peer_port);
+void libcfs_sock_release(struct socket *sock);
+
void lnet_get_tunables(void);
int lnet_peers_start_down(void);
int lnet_peer_buffer_credits(lnet_ni_t *ni);
diff --git a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
index 3bad441..be4de56 100644
--- a/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
+++ b/drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd.c
@@ -2620,7 +2620,7 @@ static kib_dev_t *kiblnd_create_dev(char *ifname)
int up;
int rc;
- rc = libcfs_ipif_query(ifname, &up, &ip, &netmask);
+ rc = lnet_ipif_query(ifname, &up, &ip, &netmask);
if (rc != 0) {
CERROR("Can't query IPoIB interface %s: %d\n",
ifname, rc);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index e642752..a27dddf 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -968,7 +968,7 @@ ksocknal_accept(lnet_ni_t *ni, struct socket *sock)
__u32 peer_ip;
int peer_port;
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
LIBCFS_ALLOC(cr, sizeof(*cr));
@@ -2594,7 +2594,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
int rc;
int n;
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Can't enumerate interfaces: %d\n", n);
return n;
@@ -2608,7 +2608,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &mask);
+ rc = lnet_ipif_query(names[i], &up, &ip, &mask);
if (rc != 0) {
CWARN("Can't get interface %s info: %d\n",
names[i], rc);
@@ -2634,7 +2634,7 @@ ksocknal_enumerate_interfaces(ksock_net_t *net)
j++;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
if (j == 0)
CERROR("Can't find any usable interfaces\n");
@@ -2796,8 +2796,7 @@ ksocknal_startup(lnet_ni_t *ni)
if (ni->ni_interfaces[i] == NULL)
break;
- rc = libcfs_ipif_query(
- ni->ni_interfaces[i], &up,
+ rc = lnet_ipif_query(ni->ni_interfaces[i], &up,
&net->ksnn_interfaces[i].ksni_ipaddr,
&net->ksnn_interfaces[i].ksni_netmask);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
index fa7ad88..352f03e 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_cb.c
@@ -1707,7 +1707,7 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
timeout = active ? *ksocknal_tunables.ksnd_timeout :
lnet_acceptor_timeout();
- rc = libcfs_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_magic, sizeof (hello->kshm_magic), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -1726,8 +1726,8 @@ ksocknal_recv_hello (lnet_ni_t *ni, ksock_conn_t *conn,
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &hello->kshm_version,
- sizeof(hello->kshm_version), timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_version,
+ sizeof(hello->kshm_version), timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
index f5e8ab0..589f0d6 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_lib-linux.c
@@ -39,9 +39,8 @@
int
ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
{
- int rc = libcfs_sock_getaddr(conn->ksnc_sock, 1,
- &conn->ksnc_ipaddr,
- &conn->ksnc_port);
+ int rc = lnet_sock_getaddr(conn->ksnc_sock, 1, &conn->ksnc_ipaddr,
+ &conn->ksnc_port);
/* Didn't need the {get,put}connsock dance to deref ksnc_sock... */
LASSERT(!conn->ksnc_closing);
@@ -51,8 +50,7 @@ ksocknal_lib_get_conn_addrs(ksock_conn_t *conn)
return rc;
}
- rc = libcfs_sock_getaddr(conn->ksnc_sock, 0,
- &conn->ksnc_myipaddr, NULL);
+ rc = lnet_sock_getaddr(conn->ksnc_sock, 0, &conn->ksnc_myipaddr, NULL);
if (rc != 0) {
CERROR("Error %d getting sock local IP\n", rc);
return rc;
@@ -436,7 +434,7 @@ ksocknal_lib_get_conn_tunables(ksock_conn_t *conn, int *txmem, int *rxmem, int *
return -ESHUTDOWN;
}
- rc = libcfs_sock_getbuf(sock, txmem, rxmem);
+ rc = lnet_sock_getbuf(sock, txmem, rxmem);
if (rc == 0) {
len = sizeof(*nagle);
rc = kernel_getsockopt(sock, SOL_TCP, TCP_NODELAY,
@@ -498,9 +496,8 @@ ksocknal_lib_setup_sock(struct socket *sock)
}
}
- rc = libcfs_sock_setbuf(sock,
- *ksocknal_tunables.ksnd_tx_buffer_size,
- *ksocknal_tunables.ksnd_rx_buffer_size);
+ rc = lnet_sock_setbuf(sock, *ksocknal_tunables.ksnd_tx_buffer_size,
+ *ksocknal_tunables.ksnd_rx_buffer_size);
if (rc != 0) {
CERROR("Can't set buffer tx %d, rx %d buffers: %d\n",
*ksocknal_tunables.ksnd_tx_buffer_size,
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
index 8596581..48197eb 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd_proto.c
@@ -495,9 +495,7 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hdr->msg.hello.type = cpu_to_le32 (hello->kshm_ctype);
hdr->msg.hello.incarnation = cpu_to_le64 (hello->kshm_src_incarnation);
- rc = libcfs_sock_write(sock, hdr, sizeof(*hdr),
- lnet_acceptor_timeout());
-
+ rc = lnet_sock_write(sock, hdr, sizeof(*hdr), lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
rc, &conn->ksnc_ipaddr, conn->ksnc_port);
@@ -511,9 +509,9 @@ ksocknal_send_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello)
hello->kshm_ips[i] = __cpu_to_le32 (hello->kshm_ips[i]);
}
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -544,9 +542,8 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
LNET_UNLOCK();
}
- rc = libcfs_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
- lnet_acceptor_timeout());
-
+ rc = lnet_sock_write(sock, hello, offsetof(ksock_hello_msg_t, kshm_ips),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO hdr to %pI4h/%d\n",
rc, &conn->ksnc_ipaddr, conn->ksnc_port);
@@ -556,9 +553,9 @@ ksocknal_send_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello)
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_write(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32),
- lnet_acceptor_timeout());
+ rc = lnet_sock_write(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32),
+ lnet_acceptor_timeout());
if (rc != 0) {
CNETERR("Error %d sending HELLO payload (%d) to %pI4h/%d\n",
rc, hello->kshm_nips,
@@ -583,9 +580,9 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
return -ENOMEM;
}
- rc = libcfs_sock_read(sock, &hdr->src_nid,
- sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
- timeout);
+ rc = lnet_sock_read(sock, &hdr->src_nid,
+ sizeof(*hdr) - offsetof(lnet_hdr_t, src_nid),
+ timeout);
if (rc != 0) {
CERROR("Error %d reading rest of HELLO hdr from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -619,8 +616,8 @@ ksocknal_recv_hello_v1(ksock_conn_t *conn, ksock_hello_msg_t *hello,
if (hello->kshm_nips == 0)
goto out;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -656,10 +653,10 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
else
conn->ksnc_flip = 1;
- rc = libcfs_sock_read(sock, &hello->kshm_src_nid,
- offsetof(ksock_hello_msg_t, kshm_ips) -
- offsetof(ksock_hello_msg_t, kshm_src_nid),
- timeout);
+ rc = lnet_sock_read(sock, &hello->kshm_src_nid,
+ offsetof(ksock_hello_msg_t, kshm_ips) -
+ offsetof(ksock_hello_msg_t, kshm_src_nid),
+ timeout);
if (rc != 0) {
CERROR("Error %d reading HELLO from %pI4h\n",
rc, &conn->ksnc_ipaddr);
@@ -687,8 +684,8 @@ ksocknal_recv_hello_v2(ksock_conn_t *conn, ksock_hello_msg_t *hello, int timeout
if (hello->kshm_nips == 0)
return 0;
- rc = libcfs_sock_read(sock, hello->kshm_ips,
- hello->kshm_nips * sizeof(__u32), timeout);
+ rc = lnet_sock_read(sock, hello->kshm_ips,
+ hello->kshm_nips * sizeof(__u32), timeout);
if (rc != 0) {
CERROR("Error %d reading IPs from ip %pI4h\n",
rc, &conn->ksnc_ipaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/acceptor.c b/drivers/staging/lustre/lnet/lnet/acceptor.c
index 8c6726a..d323899 100644
--- a/drivers/staging/lustre/lnet/lnet/acceptor.c
+++ b/drivers/staging/lustre/lnet/lnet/acceptor.c
@@ -155,9 +155,8 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
--port) {
/* Iterate through reserved ports. */
- rc = libcfs_sock_connect(&sock, &fatal,
- local_ip, port,
- peer_ip, peer_port);
+ rc = lnet_sock_connect(&sock, &fatal, local_ip, port, peer_ip,
+ peer_port);
if (rc != 0) {
if (fatal)
goto failed;
@@ -184,8 +183,7 @@ lnet_connect(struct socket **sockp, lnet_nid_t peer_nid,
lnet_net_unlock(LNET_LOCK_EX);
}
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
goto failed_sock;
@@ -220,7 +218,7 @@ lnet_accept(struct socket *sock, __u32 magic)
LASSERT(sizeof(cr) <= 16); /* not too big for the stack */
- rc = libcfs_sock_getaddr(sock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(sock, 1, &peer_ip, &peer_port);
LASSERT(rc == 0); /* we succeeded before */
if (!lnet_accept_magic(magic, LNET_PROTO_ACCEPTOR_MAGIC)) {
@@ -234,8 +232,8 @@ lnet_accept(struct socket *sock, __u32 magic)
memset(&cr, 0, sizeof(cr));
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
+ rc = lnet_sock_write(sock, &cr, sizeof(cr),
+ accept_timeout);
if (rc != 0)
CERROR("Error sending magic+version in response to LNET magic from %pI4h: %d\n",
@@ -257,9 +255,8 @@ lnet_accept(struct socket *sock, __u32 magic)
flip = (magic != LNET_PROTO_ACCEPTOR_MAGIC);
- rc = libcfs_sock_read(sock, &cr.acr_version,
- sizeof(cr.acr_version),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_version, sizeof(cr.acr_version),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request version from %pI4h\n",
rc, &peer_ip);
@@ -280,19 +277,17 @@ lnet_accept(struct socket *sock, __u32 magic)
cr.acr_magic = LNET_PROTO_ACCEPTOR_MAGIC;
cr.acr_version = LNET_PROTO_ACCEPTOR_VERSION;
- rc = libcfs_sock_write(sock, &cr, sizeof(cr),
- accept_timeout);
-
+ rc = lnet_sock_write(sock, &cr, sizeof(cr), accept_timeout);
if (rc != 0)
CERROR("Error sending magic+version in response to version %d from %pI4h: %d\n",
peer_version, &peer_ip, rc);
return -EPROTO;
}
- rc = libcfs_sock_read(sock, &cr.acr_nid,
- sizeof(cr) -
- offsetof(lnet_acceptor_connreq_t, acr_nid),
- accept_timeout);
+ rc = lnet_sock_read(sock, &cr.acr_nid,
+ sizeof(cr) -
+ offsetof(lnet_acceptor_connreq_t, acr_nid),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
rc, &peer_ip);
@@ -343,8 +338,8 @@ lnet_acceptor(void *arg)
cfs_block_allsigs();
- rc = libcfs_sock_listen(&lnet_acceptor_state.pta_sock,
- 0, accept_port, accept_backlog);
+ rc = lnet_sock_listen(&lnet_acceptor_state.pta_sock, 0, accept_port,
+ accept_backlog);
if (rc != 0) {
if (rc == -EADDRINUSE)
LCONSOLE_ERROR_MSG(0x122, "Can't start acceptor on port %d: port already in use\n",
@@ -367,7 +362,7 @@ lnet_acceptor(void *arg)
while (!lnet_acceptor_state.pta_shutdown) {
- rc = libcfs_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
+ rc = lnet_sock_accept(&newsock, lnet_acceptor_state.pta_sock);
if (rc != 0) {
if (rc != -EAGAIN) {
CWARN("Accept error %d: pausing...\n", rc);
@@ -383,7 +378,7 @@ lnet_acceptor(void *arg)
break;
}
- rc = libcfs_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
+ rc = lnet_sock_getaddr(newsock, 1, &peer_ip, &peer_port);
if (rc != 0) {
CERROR("Can't determine new connection's address\n");
goto failed;
@@ -395,8 +390,8 @@ lnet_acceptor(void *arg)
goto failed;
}
- rc = libcfs_sock_read(newsock, &magic, sizeof(magic),
- accept_timeout);
+ rc = lnet_sock_read(newsock, &magic, sizeof(magic),
+ accept_timeout);
if (rc != 0) {
CERROR("Error %d reading connection request from %pI4h\n",
rc, &peer_ip);
diff --git a/drivers/staging/lustre/lnet/lnet/config.c b/drivers/staging/lustre/lnet/lnet/config.c
index 2dc4c4a..d8c8f2e 100644
--- a/drivers/staging/lustre/lnet/lnet/config.c
+++ b/drivers/staging/lustre/lnet/lnet/config.c
@@ -1118,7 +1118,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
__u32 *ipaddrs2;
int nip;
char **ifnames;
- int nif = libcfs_ipif_enumerate(&ifnames);
+ int nif = lnet_ipif_enumerate(&ifnames);
int i;
int rc;
@@ -1128,7 +1128,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
LIBCFS_ALLOC(ipaddrs, nif * sizeof(*ipaddrs));
if (ipaddrs == NULL) {
CERROR("Can't allocate ipaddrs[%d]\n", nif);
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
return -ENOMEM;
}
@@ -1136,8 +1136,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
if (!strcmp(ifnames[i], "lo"))
continue;
- rc = libcfs_ipif_query(ifnames[i], &up,
- &ipaddrs[nip], &netmask);
+ rc = lnet_ipif_query(ifnames[i], &up, &ipaddrs[nip], &netmask);
if (rc != 0) {
CWARN("Can't query interface %s: %d\n",
ifnames[i], rc);
@@ -1153,7 +1152,7 @@ lnet_ipaddr_enumerate(__u32 **ipaddrsp)
nip++;
}
- libcfs_ipif_free_enumeration(ifnames, nif);
+ lnet_ipif_free_enumeration(ifnames, nif);
if (nip == nif) {
*ipaddrsp = ipaddrs;
@@ -1237,8 +1236,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return -EPERM;
}
- rc = libcfs_ipif_query(ni->ni_interfaces[0],
- &up, &ip, &netmask);
+ rc = lnet_ipif_query(ni->ni_interfaces[0], &up, &ip, &netmask);
if (rc != 0) {
CERROR("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), ni->ni_interfaces[0], rc);
@@ -1255,7 +1253,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
return 0;
}
- n = libcfs_ipif_enumerate(&names);
+ n = lnet_ipif_enumerate(&names);
if (n <= 0) {
CERROR("Net %s can't enumerate interfaces: %d\n",
libcfs_net2str(net), n);
@@ -1266,8 +1264,7 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
if (!strcmp(names[i], "lo")) /* skip the loopback IF */
continue;
- rc = libcfs_ipif_query(names[i], &up, &ip, &netmask);
-
+ rc = lnet_ipif_query(names[i], &up, &ip, &netmask);
if (rc != 0) {
CWARN("Net %s can't query interface %s: %d\n",
libcfs_net2str(net), names[i], rc);
@@ -1280,13 +1277,13 @@ lnet_set_ip_niaddr(lnet_ni_t *ni)
continue;
}
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
ni->ni_nid = LNET_MKNID(net, ip);
return 0;
}
CERROR("Net %s can't find any interfaces\n", libcfs_net2str(net));
- libcfs_ipif_free_enumeration(names, n);
+ lnet_ipif_free_enumeration(names, n);
return -ENOENT;
}
EXPORT_SYMBOL(lnet_set_ip_niaddr);
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index bb8d9c2..2e87168 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -45,7 +45,7 @@
#include <linux/syscalls.h>
static int
-libcfs_sock_ioctl(int cmd, unsigned long arg)
+lnet_sock_ioctl(int cmd, unsigned long arg)
{
mm_segment_t oldmm = get_fs();
struct socket *sock;
@@ -76,7 +76,7 @@ out:
}
int
-libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
+lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
{
struct ifreq ifr;
int nob;
@@ -92,8 +92,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
strcpy(ifr.ifr_name, name);
- rc = libcfs_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get flags for interface %s\n", name);
return rc;
@@ -110,8 +109,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
strcpy(ifr.ifr_name, name);
ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFADDR, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get IP address for interface %s\n", name);
return rc;
@@ -122,8 +120,7 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
strcpy(ifr.ifr_name, name);
ifr.ifr_addr.sa_family = AF_INET;
- rc = libcfs_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
-
+ rc = lnet_sock_ioctl(SIOCGIFNETMASK, (unsigned long)&ifr);
if (rc != 0) {
CERROR("Can't get netmask for interface %s\n", name);
return rc;
@@ -135,10 +132,10 @@ libcfs_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return 0;
}
-EXPORT_SYMBOL(libcfs_ipif_query);
+EXPORT_SYMBOL(lnet_ipif_query);
int
-libcfs_ipif_enumerate (char ***namesp)
+lnet_ipif_enumerate (char ***namesp)
{
/* Allocate and fill in 'names', returning # interfaces/error */
char **names;
@@ -172,8 +169,7 @@ libcfs_ipif_enumerate (char ***namesp)
ifc.ifc_buf = (char *)ifr;
ifc.ifc_len = nalloc * sizeof(*ifr);
- rc = libcfs_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
-
+ rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
if (rc < 0) {
CERROR ("Error %d enumerating interfaces\n", rc);
goto out1;
@@ -226,17 +222,17 @@ libcfs_ipif_enumerate (char ***namesp)
out2:
if (rc < 0)
- libcfs_ipif_free_enumeration(names, nfound);
+ lnet_ipif_free_enumeration(names, nfound);
out1:
LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
out0:
return rc;
}
-EXPORT_SYMBOL(libcfs_ipif_enumerate);
+EXPORT_SYMBOL(lnet_ipif_enumerate);
void
-libcfs_ipif_free_enumeration (char **names, int n)
+lnet_ipif_free_enumeration (char **names, int n)
{
int i;
@@ -248,10 +244,10 @@ libcfs_ipif_free_enumeration (char **names, int n)
LIBCFS_FREE(names, n * sizeof(*names));
}
-EXPORT_SYMBOL(libcfs_ipif_free_enumeration);
+EXPORT_SYMBOL(lnet_ipif_free_enumeration);
int
-libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
{
int rc;
long ticks = timeout * HZ;
@@ -310,10 +306,10 @@ libcfs_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_write);
+EXPORT_SYMBOL(lnet_sock_write);
int
-libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
{
int rc;
long ticks = timeout * HZ;
@@ -366,10 +362,10 @@ libcfs_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
}
}
-EXPORT_SYMBOL(libcfs_sock_read);
+EXPORT_SYMBOL(lnet_sock_read);
static int
-libcfs_sock_create (struct socket **sockp, int *fatal,
+lnet_sock_create (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port)
{
struct sockaddr_in locaddr;
@@ -424,7 +420,7 @@ libcfs_sock_create (struct socket **sockp, int *fatal,
}
int
-libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
+lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
{
int option;
int rc;
@@ -454,10 +450,10 @@ libcfs_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_setbuf);
+EXPORT_SYMBOL(lnet_sock_setbuf);
int
-libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
+lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
{
struct sockaddr_in sin;
int len = sizeof (sin);
@@ -480,10 +476,10 @@ libcfs_sock_getaddr (struct socket *sock, int remote, __u32 *ip, int *port)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_getaddr);
+EXPORT_SYMBOL(lnet_sock_getaddr);
int
-libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
+lnet_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
{
if (txbufsize != NULL) {
@@ -497,16 +493,16 @@ libcfs_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
return 0;
}
-EXPORT_SYMBOL(libcfs_sock_getbuf);
+EXPORT_SYMBOL(lnet_sock_getbuf);
int
-libcfs_sock_listen (struct socket **sockp,
+lnet_sock_listen (struct socket **sockp,
__u32 local_ip, int local_port, int backlog)
{
int fatal;
int rc;
- rc = libcfs_sock_create(sockp, &fatal, local_ip, local_port);
+ rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
if (rc != 0) {
if (!fatal)
CERROR("Can't create socket: port %d already in use\n",
@@ -523,10 +519,10 @@ libcfs_sock_listen (struct socket **sockp,
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_listen);
+EXPORT_SYMBOL(lnet_sock_listen);
int
-libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
+lnet_sock_accept (struct socket **newsockp, struct socket *sock)
{
wait_queue_t wait;
struct socket *newsock;
@@ -566,17 +562,17 @@ libcfs_sock_accept (struct socket **newsockp, struct socket *sock)
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_accept);
+EXPORT_SYMBOL(lnet_sock_accept);
int
-libcfs_sock_connect (struct socket **sockp, int *fatal,
+lnet_sock_connect (struct socket **sockp, int *fatal,
__u32 local_ip, int local_port,
__u32 peer_ip, int peer_port)
{
struct sockaddr_in srvaddr;
int rc;
- rc = libcfs_sock_create(sockp, fatal, local_ip, local_port);
+ rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
if (rc != 0)
return rc;
@@ -605,4 +601,4 @@ libcfs_sock_connect (struct socket **sockp, int *fatal,
return rc;
}
-EXPORT_SYMBOL(libcfs_sock_connect);
+EXPORT_SYMBOL(lnet_sock_connect);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 4/6] staging:lustre: use available kernel wrappers in lib-socket.c
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
Instead of handling calls to struct proto ourselves we can use
equivalent kernel wrappers. No wrapper exist for unlocked ioctl
handling so we create one here for our use. I expect some day
that function will be integrated into sock.c.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-socket.c | 47 ++++++++++++++++---------
1 files changed, 30 insertions(+), 17 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 2e87168..2d46a69 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -35,22 +35,37 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
-#include "../../include/linux/libcfs/libcfs.h"
-#include "../../include/linux/lnet/lib-lnet.h"
-
#include <linux/if.h>
#include <linux/in.h>
+#include <linux/net.h>
#include <linux/file.h>
+#include <linux/pagemap.h>
/* For sys_open & sys_close */
#include <linux/syscalls.h>
+#include <net/sock.h>
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/lib-lnet.h"
+
+static int
+kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
+{
+ mm_segment_t oldfs = get_fs();
+ int err;
+
+ set_fs(KERNEL_DS);
+ err = filp->f_op->unlocked_ioctl(filp, cmd, arg);
+ set_fs(oldfs);
+
+ return err;
+}
static int
lnet_sock_ioctl(int cmd, unsigned long arg)
{
- mm_segment_t oldmm = get_fs();
+ struct file *sock_filp;
struct socket *sock;
int rc;
- struct file *sock_filp;
rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
if (rc != 0) {
@@ -65,10 +80,7 @@ lnet_sock_ioctl(int cmd, unsigned long arg)
goto out;
}
- set_fs(KERNEL_DS);
- if (sock_filp->f_op->unlocked_ioctl)
- rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
- set_fs(oldmm);
+ rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
fput(sock_filp);
out:
@@ -398,8 +410,8 @@ lnet_sock_create (struct socket **sockp, int *fatal,
locaddr.sin_addr.s_addr = (local_ip == 0) ?
INADDR_ANY : htonl(local_ip);
- rc = sock->ops->bind(sock, (struct sockaddr *)&locaddr,
- sizeof(locaddr));
+ rc = kernel_bind(sock, (struct sockaddr *)&locaddr,
+ sizeof(locaddr));
if (rc == -EADDRINUSE) {
CDEBUG(D_NET, "Port %d already in use\n", local_port);
*fatal = 0;
@@ -459,8 +471,10 @@ lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
int len = sizeof (sin);
int rc;
- rc = sock->ops->getname (sock, (struct sockaddr *)&sin, &len,
- remote ? 2 : 0);
+ if (remote)
+ rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
+ else
+ rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
if (rc != 0) {
CERROR ("Error %d getting sock %s IP/port\n",
rc, remote ? "peer" : "local");
@@ -510,7 +524,7 @@ lnet_sock_listen (struct socket **sockp,
return rc;
}
- rc = (*sockp)->ops->listen(*sockp, backlog);
+ rc = kernel_listen(*sockp, backlog);
if (rc == 0)
return 0;
@@ -581,9 +595,8 @@ lnet_sock_connect (struct socket **sockp, int *fatal,
srvaddr.sin_port = htons(peer_port);
srvaddr.sin_addr.s_addr = htonl(peer_ip);
- rc = (*sockp)->ops->connect(*sockp,
- (struct sockaddr *)&srvaddr, sizeof(srvaddr),
- 0);
+ rc = kernel_connect(*sockp, (struct sockaddr *)&srvaddr,
+ sizeof(srvaddr), 0);
if (rc == 0)
return 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 4/6] staging:lustre: use available kernel wrappers in lib-socket.c
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
Instead of handling calls to struct proto ourselves we can use
equivalent kernel wrappers. No wrapper exist for unlocked ioctl
handling so we create one here for our use. I expect some day
that function will be integrated into sock.c.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-socket.c | 47 ++++++++++++++++---------
1 files changed, 30 insertions(+), 17 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 2e87168..2d46a69 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -35,22 +35,37 @@
*/
#define DEBUG_SUBSYSTEM S_LNET
-#include "../../include/linux/libcfs/libcfs.h"
-#include "../../include/linux/lnet/lib-lnet.h"
-
#include <linux/if.h>
#include <linux/in.h>
+#include <linux/net.h>
#include <linux/file.h>
+#include <linux/pagemap.h>
/* For sys_open & sys_close */
#include <linux/syscalls.h>
+#include <net/sock.h>
+
+#include "../../include/linux/libcfs/libcfs.h"
+#include "../../include/linux/lnet/lib-lnet.h"
+
+static int
+kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
+{
+ mm_segment_t oldfs = get_fs();
+ int err;
+
+ set_fs(KERNEL_DS);
+ err = filp->f_op->unlocked_ioctl(filp, cmd, arg);
+ set_fs(oldfs);
+
+ return err;
+}
static int
lnet_sock_ioctl(int cmd, unsigned long arg)
{
- mm_segment_t oldmm = get_fs();
+ struct file *sock_filp;
struct socket *sock;
int rc;
- struct file *sock_filp;
rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
if (rc != 0) {
@@ -65,10 +80,7 @@ lnet_sock_ioctl(int cmd, unsigned long arg)
goto out;
}
- set_fs(KERNEL_DS);
- if (sock_filp->f_op->unlocked_ioctl)
- rc = sock_filp->f_op->unlocked_ioctl(sock_filp, cmd, arg);
- set_fs(oldmm);
+ rc = kernel_sock_unlocked_ioctl(sock_filp, cmd, arg);
fput(sock_filp);
out:
@@ -398,8 +410,8 @@ lnet_sock_create (struct socket **sockp, int *fatal,
locaddr.sin_addr.s_addr = (local_ip == 0) ?
INADDR_ANY : htonl(local_ip);
- rc = sock->ops->bind(sock, (struct sockaddr *)&locaddr,
- sizeof(locaddr));
+ rc = kernel_bind(sock, (struct sockaddr *)&locaddr,
+ sizeof(locaddr));
if (rc == -EADDRINUSE) {
CDEBUG(D_NET, "Port %d already in use\n", local_port);
*fatal = 0;
@@ -459,8 +471,10 @@ lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
int len = sizeof (sin);
int rc;
- rc = sock->ops->getname (sock, (struct sockaddr *)&sin, &len,
- remote ? 2 : 0);
+ if (remote)
+ rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
+ else
+ rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
if (rc != 0) {
CERROR ("Error %d getting sock %s IP/port\n",
rc, remote ? "peer" : "local");
@@ -510,7 +524,7 @@ lnet_sock_listen (struct socket **sockp,
return rc;
}
- rc = (*sockp)->ops->listen(*sockp, backlog);
+ rc = kernel_listen(*sockp, backlog);
if (rc == 0)
return 0;
@@ -581,9 +595,8 @@ lnet_sock_connect (struct socket **sockp, int *fatal,
srvaddr.sin_port = htons(peer_port);
srvaddr.sin_addr.s_addr = htonl(peer_ip);
- rc = (*sockp)->ops->connect(*sockp,
- (struct sockaddr *)&srvaddr, sizeof(srvaddr),
- 0);
+ rc = kernel_connect(*sockp, (struct sockaddr *)&srvaddr,
+ sizeof(srvaddr), 0);
if (rc == 0)
return 0;
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
Handle all the style issues reported by checkpatch.pl.
Remove general white spaces, spaces in function calls,
etc.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 4 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 210 +++++++++-----------
2 files changed, 98 insertions(+), 116 deletions(-)
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index a27dddf..caccb00 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -1378,14 +1378,14 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
ksocknal_txlist_done(ni, &zombies, 1);
ksocknal_peer_decref(peer);
- failed_1:
+failed_1:
if (hello != NULL)
LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
kshm_ips[LNET_MAX_INTERFACES]));
LIBCFS_FREE(conn, sizeof(*conn));
- failed_0:
+failed_0:
sock_release(sock);
return rc;
}
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 2d46a69..f0b187d 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -63,13 +63,13 @@ kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
static int
lnet_sock_ioctl(int cmd, unsigned long arg)
{
- struct file *sock_filp;
- struct socket *sock;
- int rc;
+ struct file *sock_filp;
+ struct socket *sock;
+ int rc;
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
+ CERROR("Can't create socket: %d\n", rc);
return rc;
}
@@ -88,12 +88,12 @@ out:
}
int
-lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
+lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask)
{
- struct ifreq ifr;
- int nob;
- int rc;
- __u32 val;
+ struct ifreq ifr;
+ int nob;
+ int rc;
+ __u32 val;
nob = strnlen(name, IFNAMSIZ);
if (nob == IFNAMSIZ) {
@@ -101,7 +101,7 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return -EINVAL;
}
- CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
+ CLASSERT(sizeof(ifr.ifr_name) >= IFNAMSIZ);
strcpy(ifr.ifr_name, name);
rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
@@ -116,7 +116,6 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
*ip = *mask = 0;
return 0;
}
-
*up = 1;
strcpy(ifr.ifr_name, name);
@@ -143,23 +142,21 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return 0;
}
-
EXPORT_SYMBOL(lnet_ipif_query);
int
-lnet_ipif_enumerate (char ***namesp)
+lnet_ipif_enumerate(char ***namesp)
{
/* Allocate and fill in 'names', returning # interfaces/error */
- char **names;
- int toobig;
- int nalloc;
- int nfound;
- struct ifreq *ifr;
- struct ifconf ifc;
- int rc;
- int nob;
- int i;
-
+ char **names;
+ int toobig;
+ int nalloc;
+ int nfound;
+ struct ifreq *ifr;
+ struct ifconf ifc;
+ int rc;
+ int nob;
+ int i;
nalloc = 16; /* first guess at max interfaces */
toobig = 0;
@@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
toobig = 1;
nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
- CWARN("Too many interfaces: only enumerating first %d\n",
- nalloc);
+ CWARN("Too many interfaces: only enumerating "
+ "first %d\n", nalloc);
}
LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
if (ifr == NULL) {
- CERROR ("ENOMEM enumerating up to %d interfaces\n", nalloc);
+ CERROR("ENOMEM enumerating up to %d interfaces\n",
+ nalloc);
rc = -ENOMEM;
goto out0;
}
@@ -183,14 +181,14 @@ lnet_ipif_enumerate (char ***namesp)
rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
if (rc < 0) {
- CERROR ("Error %d enumerating interfaces\n", rc);
+ CERROR("Error %d enumerating interfaces\n", rc);
goto out1;
}
- LASSERT (rc == 0);
+ LASSERT(rc == 0);
nfound = ifc.ifc_len/sizeof(*ifr);
- LASSERT (nfound <= nalloc);
+ LASSERT(nfound <= nalloc);
if (nfound < nalloc || toobig)
break;
@@ -209,8 +207,7 @@ lnet_ipif_enumerate (char ***namesp)
}
for (i = 0; i < nfound; i++) {
-
- nob = strnlen (ifr[i].ifr_name, IFNAMSIZ);
+ nob = strnlen(ifr[i].ifr_name, IFNAMSIZ);
if (nob == IFNAMSIZ) {
/* no space for terminating NULL */
CERROR("interface name %.*s too long (%d max)\n",
@@ -232,41 +229,39 @@ lnet_ipif_enumerate (char ***namesp)
*namesp = names;
rc = nfound;
- out2:
+out2:
if (rc < 0)
lnet_ipif_free_enumeration(names, nfound);
- out1:
+out1:
LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- out0:
+out0:
return rc;
}
-
EXPORT_SYMBOL(lnet_ipif_enumerate);
void
-lnet_ipif_free_enumeration (char **names, int n)
+lnet_ipif_free_enumeration(char **names, int n)
{
- int i;
+ int i;
- LASSERT (n > 0);
+ LASSERT(n > 0);
for (i = 0; i < n && names[i] != NULL; i++)
LIBCFS_FREE(names[i], IFNAMSIZ);
LIBCFS_FREE(names, n * sizeof(*names));
}
-
EXPORT_SYMBOL(lnet_ipif_free_enumeration);
int
-lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout)
{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
+ int rc;
+ long ticks = timeout * HZ;
+ unsigned long then;
struct timeval tv;
- LASSERT (nob > 0);
+ LASSERT(nob > 0);
/* Caller may pass a zero timeout if she thinks the socket buffer is
* empty enough to take the whole message immediately */
@@ -286,9 +281,10 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
.tv_usec = ((ticks % HZ) * 1000000) / HZ
};
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
- (char *)&tv, sizeof(tv));
+ (char *)&tv, sizeof(tv));
if (rc != 0) {
- CERROR("Can't set socket send timeout %ld.%06d: %d\n",
+ CERROR("Can't set socket send timeout "
+ "%ld.%06d: %d\n",
(long)tv.tv_sec, (int)tv.tv_usec, rc);
return rc;
}
@@ -305,7 +301,7 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
return rc;
if (rc == 0) {
- CERROR ("Unexpected zero rc\n");
+ CERROR("Unexpected zero rc\n");
return -ECONNABORTED;
}
@@ -315,21 +311,20 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
buffer = ((char *)buffer) + rc;
nob -= rc;
}
-
return 0;
}
EXPORT_SYMBOL(lnet_sock_write);
int
-lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout)
{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
+ int rc;
+ long ticks = timeout * HZ;
+ unsigned long then;
struct timeval tv;
- LASSERT (nob > 0);
- LASSERT (ticks > 0);
+ LASSERT(nob > 0);
+ LASSERT(ticks > 0);
for (;;) {
struct kvec iov = {
@@ -337,7 +332,7 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
.iov_len = nob
};
struct msghdr msg = {
- .msg_flags = 0
+ .msg_flags = 0
};
/* Set receive timeout to remaining time */
@@ -346,7 +341,7 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
.tv_usec = ((ticks % HZ) * 1000000) / HZ
};
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
- (char *)&tv, sizeof(tv));
+ (char *)&tv, sizeof(tv));
if (rc != 0) {
CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
(long)tv.tv_sec, (int)tv.tv_usec, rc);
@@ -373,31 +368,30 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
return -ETIMEDOUT;
}
}
-
EXPORT_SYMBOL(lnet_sock_read);
static int
-lnet_sock_create (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port)
+lnet_sock_create(struct socket **sockp, int *fatal, __u32 local_ip,
+ int local_port)
{
- struct sockaddr_in locaddr;
- struct socket *sock;
- int rc;
- int option;
+ struct sockaddr_in locaddr;
+ struct socket *sock;
+ int rc;
+ int option;
/* All errors are fatal except bind failure if the port is in use */
*fatal = 1;
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
*sockp = sock;
if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
+ CERROR("Can't create socket: %d\n", rc);
return rc;
}
option = 1;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
goto failed;
@@ -423,27 +417,26 @@ lnet_sock_create (struct socket **sockp, int *fatal,
goto failed;
}
}
-
return 0;
- failed:
+failed:
sock_release(sock);
return rc;
}
int
-lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
+lnet_sock_setbuf(struct socket *sock, int txbufsize, int rxbufsize)
{
- int option;
- int rc;
+ int option;
+ int rc;
if (txbufsize != 0) {
option = txbufsize;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
- CERROR ("Can't set send buffer %d: %d\n",
- option, rc);
+ CERROR("Can't set send buffer %d: %d\n",
+ option, rc);
return rc;
}
}
@@ -451,70 +444,63 @@ lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
if (rxbufsize != 0) {
option = rxbufsize;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
- CERROR ("Can't set receive buffer %d: %d\n",
- option, rc);
+ CERROR("Can't set receive buffer %d: %d\n",
+ option, rc);
return rc;
}
}
-
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_setbuf);
int
-lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
+lnet_sock_getaddr(struct socket *sock, bool remote, __u32 *ip, int *port)
{
struct sockaddr_in sin;
- int len = sizeof (sin);
- int rc;
+ int len = sizeof(sin);
+ int rc;
if (remote)
rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
else
rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
if (rc != 0) {
- CERROR ("Error %d getting sock %s IP/port\n",
- rc, remote ? "peer" : "local");
+ CERROR("Error %d getting sock %s IP/port\n",
+ rc, remote ? "peer" : "local");
return rc;
}
if (ip != NULL)
- *ip = ntohl (sin.sin_addr.s_addr);
+ *ip = ntohl(sin.sin_addr.s_addr);
if (port != NULL)
- *port = ntohs (sin.sin_port);
+ *port = ntohs(sin.sin_port);
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_getaddr);
int
-lnet_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
+lnet_sock_getbuf(struct socket *sock, int *txbufsize, int *rxbufsize)
{
-
- if (txbufsize != NULL) {
+ if (txbufsize != NULL)
*txbufsize = sock->sk->sk_sndbuf;
- }
- if (rxbufsize != NULL) {
+ if (rxbufsize != NULL)
*rxbufsize = sock->sk->sk_rcvbuf;
- }
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_getbuf);
int
-lnet_sock_listen (struct socket **sockp,
- __u32 local_ip, int local_port, int backlog)
+lnet_sock_listen(struct socket **sockp, __u32 local_ip, int local_port,
+ int backlog)
{
- int fatal;
- int rc;
+ int fatal;
+ int rc;
rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
if (rc != 0) {
@@ -532,15 +518,14 @@ lnet_sock_listen (struct socket **sockp,
sock_release(*sockp);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_listen);
int
-lnet_sock_accept (struct socket **newsockp, struct socket *sock)
+lnet_sock_accept(struct socket **newsockp, struct socket *sock)
{
- wait_queue_t wait;
+ wait_queue_t wait;
struct socket *newsock;
- int rc;
+ int rc;
init_waitqueue_entry(&wait, current);
@@ -571,26 +556,24 @@ lnet_sock_accept (struct socket **newsockp, struct socket *sock)
*newsockp = newsock;
return 0;
- failed:
+failed:
sock_release(newsock);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_accept);
int
-lnet_sock_connect (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port)
+lnet_sock_connect(struct socket **sockp, int *fatal, __u32 local_ip,
+ int local_port, __u32 peer_ip, int peer_port)
{
- struct sockaddr_in srvaddr;
- int rc;
+ struct sockaddr_in srvaddr;
+ int rc;
rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
if (rc != 0)
return rc;
- memset (&srvaddr, 0, sizeof (srvaddr));
+ memset(&srvaddr, 0, sizeof(srvaddr));
srvaddr.sin_family = AF_INET;
srvaddr.sin_port = htons(peer_port);
srvaddr.sin_addr.s_addr = htonl(peer_ip);
@@ -607,11 +590,10 @@ lnet_sock_connect (struct socket **sockp, int *fatal,
*fatal = !(rc == -EADDRNOTAVAIL);
CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
- "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
- &local_ip, local_port, &peer_ip, peer_port);
+ "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
+ &local_ip, local_port, &peer_ip, peer_port);
sock_release(*sockp);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_connect);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
Handle all the style issues reported by checkpatch.pl.
Remove general white spaces, spaces in function calls,
etc.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../staging/lustre/lnet/klnds/socklnd/socklnd.c | 4 +-
drivers/staging/lustre/lnet/lnet/lib-socket.c | 210 +++++++++-----------
2 files changed, 98 insertions(+), 116 deletions(-)
diff --git a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
index a27dddf..caccb00 100644
--- a/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
+++ b/drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c
@@ -1378,14 +1378,14 @@ ksocknal_create_conn(lnet_ni_t *ni, ksock_route_t *route,
ksocknal_txlist_done(ni, &zombies, 1);
ksocknal_peer_decref(peer);
- failed_1:
+failed_1:
if (hello != NULL)
LIBCFS_FREE(hello, offsetof(ksock_hello_msg_t,
kshm_ips[LNET_MAX_INTERFACES]));
LIBCFS_FREE(conn, sizeof(*conn));
- failed_0:
+failed_0:
sock_release(sock);
return rc;
}
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index 2d46a69..f0b187d 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -63,13 +63,13 @@ kernel_sock_unlocked_ioctl(struct file *filp, int cmd, unsigned long arg)
static int
lnet_sock_ioctl(int cmd, unsigned long arg)
{
- struct file *sock_filp;
- struct socket *sock;
- int rc;
+ struct file *sock_filp;
+ struct socket *sock;
+ int rc;
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
+ CERROR("Can't create socket: %d\n", rc);
return rc;
}
@@ -88,12 +88,12 @@ out:
}
int
-lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
+lnet_ipif_query(char *name, int *up, __u32 *ip, __u32 *mask)
{
- struct ifreq ifr;
- int nob;
- int rc;
- __u32 val;
+ struct ifreq ifr;
+ int nob;
+ int rc;
+ __u32 val;
nob = strnlen(name, IFNAMSIZ);
if (nob == IFNAMSIZ) {
@@ -101,7 +101,7 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return -EINVAL;
}
- CLASSERT (sizeof(ifr.ifr_name) >= IFNAMSIZ);
+ CLASSERT(sizeof(ifr.ifr_name) >= IFNAMSIZ);
strcpy(ifr.ifr_name, name);
rc = lnet_sock_ioctl(SIOCGIFFLAGS, (unsigned long)&ifr);
@@ -116,7 +116,6 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
*ip = *mask = 0;
return 0;
}
-
*up = 1;
strcpy(ifr.ifr_name, name);
@@ -143,23 +142,21 @@ lnet_ipif_query (char *name, int *up, __u32 *ip, __u32 *mask)
return 0;
}
-
EXPORT_SYMBOL(lnet_ipif_query);
int
-lnet_ipif_enumerate (char ***namesp)
+lnet_ipif_enumerate(char ***namesp)
{
/* Allocate and fill in 'names', returning # interfaces/error */
- char **names;
- int toobig;
- int nalloc;
- int nfound;
- struct ifreq *ifr;
- struct ifconf ifc;
- int rc;
- int nob;
- int i;
-
+ char **names;
+ int toobig;
+ int nalloc;
+ int nfound;
+ struct ifreq *ifr;
+ struct ifconf ifc;
+ int rc;
+ int nob;
+ int i;
nalloc = 16; /* first guess at max interfaces */
toobig = 0;
@@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
toobig = 1;
nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
- CWARN("Too many interfaces: only enumerating first %d\n",
- nalloc);
+ CWARN("Too many interfaces: only enumerating "
+ "first %d\n", nalloc);
}
LIBCFS_ALLOC(ifr, nalloc * sizeof(*ifr));
if (ifr == NULL) {
- CERROR ("ENOMEM enumerating up to %d interfaces\n", nalloc);
+ CERROR("ENOMEM enumerating up to %d interfaces\n",
+ nalloc);
rc = -ENOMEM;
goto out0;
}
@@ -183,14 +181,14 @@ lnet_ipif_enumerate (char ***namesp)
rc = lnet_sock_ioctl(SIOCGIFCONF, (unsigned long)&ifc);
if (rc < 0) {
- CERROR ("Error %d enumerating interfaces\n", rc);
+ CERROR("Error %d enumerating interfaces\n", rc);
goto out1;
}
- LASSERT (rc == 0);
+ LASSERT(rc == 0);
nfound = ifc.ifc_len/sizeof(*ifr);
- LASSERT (nfound <= nalloc);
+ LASSERT(nfound <= nalloc);
if (nfound < nalloc || toobig)
break;
@@ -209,8 +207,7 @@ lnet_ipif_enumerate (char ***namesp)
}
for (i = 0; i < nfound; i++) {
-
- nob = strnlen (ifr[i].ifr_name, IFNAMSIZ);
+ nob = strnlen(ifr[i].ifr_name, IFNAMSIZ);
if (nob == IFNAMSIZ) {
/* no space for terminating NULL */
CERROR("interface name %.*s too long (%d max)\n",
@@ -232,41 +229,39 @@ lnet_ipif_enumerate (char ***namesp)
*namesp = names;
rc = nfound;
- out2:
+out2:
if (rc < 0)
lnet_ipif_free_enumeration(names, nfound);
- out1:
+out1:
LIBCFS_FREE(ifr, nalloc * sizeof(*ifr));
- out0:
+out0:
return rc;
}
-
EXPORT_SYMBOL(lnet_ipif_enumerate);
void
-lnet_ipif_free_enumeration (char **names, int n)
+lnet_ipif_free_enumeration(char **names, int n)
{
- int i;
+ int i;
- LASSERT (n > 0);
+ LASSERT(n > 0);
for (i = 0; i < n && names[i] != NULL; i++)
LIBCFS_FREE(names[i], IFNAMSIZ);
LIBCFS_FREE(names, n * sizeof(*names));
}
-
EXPORT_SYMBOL(lnet_ipif_free_enumeration);
int
-lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_write(struct socket *sock, void *buffer, int nob, int timeout)
{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
+ int rc;
+ long ticks = timeout * HZ;
+ unsigned long then;
struct timeval tv;
- LASSERT (nob > 0);
+ LASSERT(nob > 0);
/* Caller may pass a zero timeout if she thinks the socket buffer is
* empty enough to take the whole message immediately */
@@ -286,9 +281,10 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
.tv_usec = ((ticks % HZ) * 1000000) / HZ
};
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO,
- (char *)&tv, sizeof(tv));
+ (char *)&tv, sizeof(tv));
if (rc != 0) {
- CERROR("Can't set socket send timeout %ld.%06d: %d\n",
+ CERROR("Can't set socket send timeout "
+ "%ld.%06d: %d\n",
(long)tv.tv_sec, (int)tv.tv_usec, rc);
return rc;
}
@@ -305,7 +301,7 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
return rc;
if (rc == 0) {
- CERROR ("Unexpected zero rc\n");
+ CERROR("Unexpected zero rc\n");
return -ECONNABORTED;
}
@@ -315,21 +311,20 @@ lnet_sock_write (struct socket *sock, void *buffer, int nob, int timeout)
buffer = ((char *)buffer) + rc;
nob -= rc;
}
-
return 0;
}
EXPORT_SYMBOL(lnet_sock_write);
int
-lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
+lnet_sock_read(struct socket *sock, void *buffer, int nob, int timeout)
{
- int rc;
- long ticks = timeout * HZ;
- unsigned long then;
+ int rc;
+ long ticks = timeout * HZ;
+ unsigned long then;
struct timeval tv;
- LASSERT (nob > 0);
- LASSERT (ticks > 0);
+ LASSERT(nob > 0);
+ LASSERT(ticks > 0);
for (;;) {
struct kvec iov = {
@@ -337,7 +332,7 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
.iov_len = nob
};
struct msghdr msg = {
- .msg_flags = 0
+ .msg_flags = 0
};
/* Set receive timeout to remaining time */
@@ -346,7 +341,7 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
.tv_usec = ((ticks % HZ) * 1000000) / HZ
};
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO,
- (char *)&tv, sizeof(tv));
+ (char *)&tv, sizeof(tv));
if (rc != 0) {
CERROR("Can't set socket recv timeout %ld.%06d: %d\n",
(long)tv.tv_sec, (int)tv.tv_usec, rc);
@@ -373,31 +368,30 @@ lnet_sock_read (struct socket *sock, void *buffer, int nob, int timeout)
return -ETIMEDOUT;
}
}
-
EXPORT_SYMBOL(lnet_sock_read);
static int
-lnet_sock_create (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port)
+lnet_sock_create(struct socket **sockp, int *fatal, __u32 local_ip,
+ int local_port)
{
- struct sockaddr_in locaddr;
- struct socket *sock;
- int rc;
- int option;
+ struct sockaddr_in locaddr;
+ struct socket *sock;
+ int rc;
+ int option;
/* All errors are fatal except bind failure if the port is in use */
*fatal = 1;
- rc = sock_create (PF_INET, SOCK_STREAM, 0, &sock);
+ rc = sock_create(PF_INET, SOCK_STREAM, 0, &sock);
*sockp = sock;
if (rc != 0) {
- CERROR ("Can't create socket: %d\n", rc);
+ CERROR("Can't create socket: %d\n", rc);
return rc;
}
option = 1;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
CERROR("Can't set SO_REUSEADDR for socket: %d\n", rc);
goto failed;
@@ -423,27 +417,26 @@ lnet_sock_create (struct socket **sockp, int *fatal,
goto failed;
}
}
-
return 0;
- failed:
+failed:
sock_release(sock);
return rc;
}
int
-lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
+lnet_sock_setbuf(struct socket *sock, int txbufsize, int rxbufsize)
{
- int option;
- int rc;
+ int option;
+ int rc;
if (txbufsize != 0) {
option = txbufsize;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_SNDBUF,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
- CERROR ("Can't set send buffer %d: %d\n",
- option, rc);
+ CERROR("Can't set send buffer %d: %d\n",
+ option, rc);
return rc;
}
}
@@ -451,70 +444,63 @@ lnet_sock_setbuf (struct socket *sock, int txbufsize, int rxbufsize)
if (rxbufsize != 0) {
option = rxbufsize;
rc = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUF,
- (char *)&option, sizeof (option));
+ (char *)&option, sizeof(option));
if (rc != 0) {
- CERROR ("Can't set receive buffer %d: %d\n",
- option, rc);
+ CERROR("Can't set receive buffer %d: %d\n",
+ option, rc);
return rc;
}
}
-
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_setbuf);
int
-lnet_sock_getaddr (struct socket *sock, bool remote, __u32 *ip, int *port)
+lnet_sock_getaddr(struct socket *sock, bool remote, __u32 *ip, int *port)
{
struct sockaddr_in sin;
- int len = sizeof (sin);
- int rc;
+ int len = sizeof(sin);
+ int rc;
if (remote)
rc = kernel_getpeername(sock, (struct sockaddr *)&sin, &len);
else
rc = kernel_getsockname(sock, (struct sockaddr *)&sin, &len);
if (rc != 0) {
- CERROR ("Error %d getting sock %s IP/port\n",
- rc, remote ? "peer" : "local");
+ CERROR("Error %d getting sock %s IP/port\n",
+ rc, remote ? "peer" : "local");
return rc;
}
if (ip != NULL)
- *ip = ntohl (sin.sin_addr.s_addr);
+ *ip = ntohl(sin.sin_addr.s_addr);
if (port != NULL)
- *port = ntohs (sin.sin_port);
+ *port = ntohs(sin.sin_port);
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_getaddr);
int
-lnet_sock_getbuf (struct socket *sock, int *txbufsize, int *rxbufsize)
+lnet_sock_getbuf(struct socket *sock, int *txbufsize, int *rxbufsize)
{
-
- if (txbufsize != NULL) {
+ if (txbufsize != NULL)
*txbufsize = sock->sk->sk_sndbuf;
- }
- if (rxbufsize != NULL) {
+ if (rxbufsize != NULL)
*rxbufsize = sock->sk->sk_rcvbuf;
- }
return 0;
}
-
EXPORT_SYMBOL(lnet_sock_getbuf);
int
-lnet_sock_listen (struct socket **sockp,
- __u32 local_ip, int local_port, int backlog)
+lnet_sock_listen(struct socket **sockp, __u32 local_ip, int local_port,
+ int backlog)
{
- int fatal;
- int rc;
+ int fatal;
+ int rc;
rc = lnet_sock_create(sockp, &fatal, local_ip, local_port);
if (rc != 0) {
@@ -532,15 +518,14 @@ lnet_sock_listen (struct socket **sockp,
sock_release(*sockp);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_listen);
int
-lnet_sock_accept (struct socket **newsockp, struct socket *sock)
+lnet_sock_accept(struct socket **newsockp, struct socket *sock)
{
- wait_queue_t wait;
+ wait_queue_t wait;
struct socket *newsock;
- int rc;
+ int rc;
init_waitqueue_entry(&wait, current);
@@ -571,26 +556,24 @@ lnet_sock_accept (struct socket **newsockp, struct socket *sock)
*newsockp = newsock;
return 0;
- failed:
+failed:
sock_release(newsock);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_accept);
int
-lnet_sock_connect (struct socket **sockp, int *fatal,
- __u32 local_ip, int local_port,
- __u32 peer_ip, int peer_port)
+lnet_sock_connect(struct socket **sockp, int *fatal, __u32 local_ip,
+ int local_port, __u32 peer_ip, int peer_port)
{
- struct sockaddr_in srvaddr;
- int rc;
+ struct sockaddr_in srvaddr;
+ int rc;
rc = lnet_sock_create(sockp, fatal, local_ip, local_port);
if (rc != 0)
return rc;
- memset (&srvaddr, 0, sizeof (srvaddr));
+ memset(&srvaddr, 0, sizeof(srvaddr));
srvaddr.sin_family = AF_INET;
srvaddr.sin_port = htons(peer_port);
srvaddr.sin_addr.s_addr = htonl(peer_ip);
@@ -607,11 +590,10 @@ lnet_sock_connect (struct socket **sockp, int *fatal,
*fatal = !(rc == -EADDRNOTAVAIL);
CDEBUG_LIMIT(*fatal ? D_NETERROR : D_NET,
- "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
- &local_ip, local_port, &peer_ip, peer_port);
+ "Error %d connecting %pI4h/%d -> %pI4h/%d\n", rc,
+ &local_ip, local_port, &peer_ip, peer_port);
sock_release(*sockp);
return rc;
}
-
EXPORT_SYMBOL(lnet_sock_connect);
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
2015-05-22 18:32 ` James Simmons
@ 2015-05-25 9:37 ` Dan Carpenter
-1 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-25 9:37 UTC (permalink / raw)
To: lustre-devel
On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
> toobig = 1;
> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
> - CWARN("Too many interfaces: only enumerating first %d\n",
> - nalloc);
> + CWARN("Too many interfaces: only enumerating "
> + "first %d\n", nalloc);
> }
Don't split string literals, it makes them hard to grep for.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* Re: [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
@ 2015-05-25 9:37 ` Dan Carpenter
0 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-25 9:37 UTC (permalink / raw)
To: James Simmons
Cc: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger,
HPDD-discuss, Linux Kernel Mailing List, lustre-devel
On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
> toobig = 1;
> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
> - CWARN("Too many interfaces: only enumerating first %d\n",
> - nalloc);
> + CWARN("Too many interfaces: only enumerating "
> + "first %d\n", nalloc);
> }
Don't split string literals, it makes them hard to grep for.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
2015-05-25 9:37 ` Dan Carpenter
@ 2015-05-27 15:01 ` Simmons, James A.
-1 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-27 15:01 UTC (permalink / raw)
To: lustre-devel
>>On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
>> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
>> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
>> toobig = 1;
>> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
>> - CWARN("Too many interfaces: only enumerating first %d\n",
>> - nalloc);
>> + CWARN("Too many interfaces: only enumerating "
>> + "first %d\n", nalloc);
>> }
>
>Don't split string literals, it makes them hard to grep for.
Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
If this is the only problem then this patch set it ready. I have more patch series that are dependent
on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
^ permalink raw reply [flat|nested] 44+ messages in thread* RE: [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
@ 2015-05-27 15:01 ` Simmons, James A.
0 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-27 15:01 UTC (permalink / raw)
To: 'Dan Carpenter', James Simmons
Cc: devel@driverdev.osuosl.org, Greg Kroah-Hartman,
Linux Kernel Mailing List, Oleg Drokin, HPDD-discuss@ml01.01.org,
lustre-devel@lists.lustre.org
>>On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
>> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
>> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
>> toobig = 1;
>> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
>> - CWARN("Too many interfaces: only enumerating first %d\n",
>> - nalloc);
>> + CWARN("Too many interfaces: only enumerating "
>> + "first %d\n", nalloc);
>> }
>
>Don't split string literals, it makes them hard to grep for.
Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
If this is the only problem then this patch set it ready. I have more patch series that are dependent
on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
2015-05-27 15:01 ` Simmons, James A.
@ 2015-05-27 15:24 ` Dan Carpenter
-1 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-27 15:24 UTC (permalink / raw)
To: lustre-devel
On Wed, May 27, 2015 at 03:01:37PM +0000, Simmons, James A. wrote:
> >>On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
> >> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
> >> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
> >> toobig = 1;
> >> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
> >> - CWARN("Too many interfaces: only enumerating first %d\n",
> >> - nalloc);
> >> + CWARN("Too many interfaces: only enumerating "
> >> + "first %d\n", nalloc);
> >> }
> >
> >Don't split string literals, it makes them hard to grep for.
>
> Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
> If this is the only problem then this patch set it ready.
Normally the right thing to do here would be to send a fixed
[patch 5/6 v2] using the --in-reply-to option so that it appears as a
reply to the original [patch 5/6].
> I have more patch series that are dependent
> on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
> cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
You will get an email when these are merged.
This is the only issue, I had. No one else has complained so that means
no one else has any objections. Greg hasn't merged it yet and he might
find a problem with it, but it seems like a straight forward patchset
so that's unlikely.
The only issue is that this patchset was sent in a confusing way. It
doesn't have a v2 tag and it was tacked on to the old thread. Greg
tends to not waste time being confused and just deletes the whole thread
when that happens.
The notes that "This depends on XXX being applied." are kind of nice but
no one spends a lot of time worrying about that stuff. Greg just
applies the patches in the order that they hit his inbox and they either
apply or he tells people to update and resend.
So send your follow on patches. If everything applies then good. If
not then you can resend, which is a few commands in git and not a big
deal.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* Re: [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
@ 2015-05-27 15:24 ` Dan Carpenter
0 siblings, 0 replies; 44+ messages in thread
From: Dan Carpenter @ 2015-05-27 15:24 UTC (permalink / raw)
To: Simmons, James A.
Cc: James Simmons, devel@driverdev.osuosl.org, Greg Kroah-Hartman,
Linux Kernel Mailing List, Oleg Drokin, HPDD-discuss@ml01.01.org,
lustre-devel@lists.lustre.org
On Wed, May 27, 2015 at 03:01:37PM +0000, Simmons, James A. wrote:
> >>On Fri, May 22, 2015 at 02:32:31PM -0400, James Simmons wrote:
> >> @@ -167,13 +164,14 @@ lnet_ipif_enumerate (char ***namesp)
> >> if (nalloc * sizeof(*ifr) > PAGE_CACHE_SIZE) {
> >> toobig = 1;
> >> nalloc = PAGE_CACHE_SIZE/sizeof(*ifr);
> >> - CWARN("Too many interfaces: only enumerating first %d\n",
> >> - nalloc);
> >> + CWARN("Too many interfaces: only enumerating "
> >> + "first %d\n", nalloc);
> >> }
> >
> >Don't split string literals, it makes them hard to grep for.
>
> Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
> If this is the only problem then this patch set it ready.
Normally the right thing to do here would be to send a fixed
[patch 5/6 v2] using the --in-reply-to option so that it appears as a
reply to the original [patch 5/6].
> I have more patch series that are dependent
> on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
> cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
You will get an email when these are merged.
This is the only issue, I had. No one else has complained so that means
no one else has any objections. Greg hasn't merged it yet and he might
find a problem with it, but it seems like a straight forward patchset
so that's unlikely.
The only issue is that this patchset was sent in a confusing way. It
doesn't have a v2 tag and it was tacked on to the old thread. Greg
tends to not waste time being confused and just deletes the whole thread
when that happens.
The notes that "This depends on XXX being applied." are kind of nice but
no one spends a lot of time worrying about that stuff. Greg just
applies the patches in the order that they hit his inbox and they either
apply or he tells people to update and resend.
So send your follow on patches. If everything applies then good. If
not then you can resend, which is a few commands in git and not a big
deal.
regards,
dan carpenter
^ permalink raw reply [flat|nested] 44+ messages in thread* [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
2015-05-27 15:24 ` Dan Carpenter
@ 2015-05-27 21:04 ` Simmons, James A.
-1 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-27 21:04 UTC (permalink / raw)
To: lustre-devel
>> >
>> >Don't split string literals, it makes them hard to grep for.
>>
>> Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
>> If this is the only problem then this patch set it ready.
>
>Normally the right thing to do here would be to send a fixed
>[patch 5/6 v2] using the --in-reply-to option so that it appears as a
>reply to the original [patch 5/6].
Made a note of that for future reference.
>> I have more patch series that are dependent
>> on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
>> cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
>
>You will get an email when these are merged.
>
>This is the only issue, I had. No one else has complained so that means
>no one else has any objections. Greg hasn't merged it yet and he might
>find a problem with it, but it seems like a straight forward patchset
>so that's unlikely.
Ugh. I was off by one for the number of patches so I need to send a new batch.
>The only issue is that this patchset was sent in a confusing way. It
>doesn't have a v2 tag and it was tacked on to the old thread. Greg
>tends to not waste time being confused and just deletes the whole thread
>when that happens.
Forgot to change the patch tag number. I will send a new batch with v2 so it
is a new thread.
^ permalink raw reply [flat|nested] 44+ messages in thread
* RE: [lustre-devel] [PATCH 5/6] staging:lustre: style cleanups for lib-socket.c
@ 2015-05-27 21:04 ` Simmons, James A.
0 siblings, 0 replies; 44+ messages in thread
From: Simmons, James A. @ 2015-05-27 21:04 UTC (permalink / raw)
To: 'Dan Carpenter'
Cc: James Simmons, devel@driverdev.osuosl.org, Greg Kroah-Hartman,
Linux Kernel Mailing List, Oleg Drokin, HPDD-discuss@ml01.01.org,
lustre-devel@lists.lustre.org
>> >
>> >Don't split string literals, it makes them hard to grep for.
>>
>> Will fix. The CWARN will go over 80 characters but from the recent emails that is more acceptable.
>> If this is the only problem then this patch set it ready.
>
>Normally the right thing to do here would be to send a fixed
>[patch 5/6 v2] using the --in-reply-to option so that it appears as a
>reply to the original [patch 5/6].
Made a note of that for future reference.
>> I have more patch series that are dependent
>> on this first one. Should I push the other patch series with a note that it is dependent on the tcpip
>> cleanup or wait until it is merged? Also how does one find out when the patch has been merged?
>
>You will get an email when these are merged.
>
>This is the only issue, I had. No one else has complained so that means
>no one else has any objections. Greg hasn't merged it yet and he might
>find a problem with it, but it seems like a straight forward patchset
>so that's unlikely.
Ugh. I was off by one for the number of patches so I need to send a new batch.
>The only issue is that this patchset was sent in a confusing way. It
>doesn't have a v2 tag and it was tacked on to the old thread. Greg
>tends to not waste time being confused and just deletes the whole thread
>when that happens.
Forgot to change the patch tag number. I will send a new batch with v2 so it
is a new thread.
^ permalink raw reply [flat|nested] 44+ messages in thread
* [lustre-devel] [PATCH 6/6] staging:lustre: Update license and copyright for lib-socket.c
[not found] <uja.ornl@gmail.com>
@ 2015-05-22 18:32 ` James Simmons
2015-05-21 22:46 ` James Simmons
` (9 subsequent siblings)
10 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: lustre-devel
Point to the right place for GNU license. Update Intel copyright.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-socket.c | 10 +++-------
1 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index f0b187d..2f8443c 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,11 +23,11 @@
* Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2015 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#define DEBUG_SUBSYSTEM S_LNET
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread* [PATCH 6/6] staging:lustre: Update license and copyright for lib-socket.c
@ 2015-05-22 18:32 ` James Simmons
0 siblings, 0 replies; 44+ messages in thread
From: James Simmons @ 2015-05-22 18:32 UTC (permalink / raw)
To: Greg Kroah-Hartman, devel, Oleg Drokin, Andreas Dilger
Cc: Linux Kernel Mailing List, HPDD-discuss, lustre-devel,
James Simmons
Point to the right place for GNU license. Update Intel copyright.
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
drivers/staging/lustre/lnet/lnet/lib-socket.c | 10 +++-------
1 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/lustre/lnet/lnet/lib-socket.c b/drivers/staging/lustre/lnet/lnet/lib-socket.c
index f0b187d..2f8443c 100644
--- a/drivers/staging/lustre/lnet/lnet/lib-socket.c
+++ b/drivers/staging/lustre/lnet/lnet/lib-socket.c
@@ -15,11 +15,7 @@
*
* You should have received a copy of the GNU General Public License
* version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
*
* GPL HEADER END
*/
@@ -27,11 +23,11 @@
* Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2012, Intel Corporation.
+ * Copyright (c) 2012, 2015 Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
+ * Lustre is a trademark of Seagate, Inc.
*/
#define DEBUG_SUBSYSTEM S_LNET
--
1.7.1
^ permalink raw reply related [flat|nested] 44+ messages in thread